{
 "schema_version": "1.0.0",
 "source": "reconstructed from SEARCH.html",
 "legacy_generated": "2026-07-19",
 "vocabularies": {
  "disciplines": [
   "anthropology",
   "art-history",
   "artificial-intelligence",
   "astronomy",
   "behavioral-economics",
   "biology",
   "chemistry",
   "cognitive-psychology",
   "cognitive-science",
   "computer-science",
   "control-theory",
   "digital-humanities",
   "earth-science",
   "ecology",
   "economics",
   "education",
   "epidemiology",
   "genetics",
   "information-theory",
   "law",
   "linguistics",
   "literature",
   "marketing",
   "materials-science",
   "mathematics",
   "medicine",
   "network-science",
   "neuroscience",
   "operations-research",
   "philosophy",
   "physics",
   "political-science",
   "psychology",
   "quantitative-finance",
   "quantum-computing",
   "robotics",
   "social-psychology",
   "sociology",
   "statistics"
  ],
  "discipline_colors": {
   "anthropology": "#0b4f9c",
   "art-history": "#1a7a33",
   "artificial-intelligence": "#8324c4",
   "astronomy": "#c0392b",
   "behavioral-economics": "#00838f",
   "biology": "#a35b00",
   "chemistry": "#b5179e",
   "cognitive-psychology": "#2e7d32",
   "cognitive-science": "#1565c0",
   "computer-science": "#6a1b9a",
   "control-theory": "#ad1457",
   "digital-humanities": "#00695c",
   "earth-science": "#5d4037",
   "ecology": "#283593",
   "economics": "#558b2f",
   "education": "#d84315",
   "epidemiology": "#0277bd",
   "genetics": "#7b1fa2",
   "information-theory": "#c62828",
   "law": "#00897b",
   "linguistics": "#4527a0",
   "literature": "#ef6c00",
   "marketing": "#3949ab",
   "materials-science": "#546e7a",
   "mathematics": "#00acc1",
   "medicine": "#9e9d24",
   "network-science": "#8d6e63",
   "neuroscience": "#5e35b1",
   "operations-research": "#43a047",
   "philosophy": "#3d5afe",
   "physics": "#0b4f9c",
   "political-science": "#1a7a33",
   "psychology": "#8324c4",
   "quantitative-finance": "#c0392b",
   "quantum-computing": "#00838f",
   "robotics": "#a35b00",
   "social-psychology": "#b5179e",
   "sociology": "#2e7d32",
   "statistics": "#1565c0"
  },
  "statistics": [
   "ANOVA",
   "Bayesian",
   "R-squared",
   "bootstrap",
   "chi-square",
   "confidence-interval",
   "correlation",
   "degrees-of-freedom",
   "effect-size",
   "false-discovery-rate",
   "hypothesis-testing",
   "mean",
   "median",
   "null-hypothesis",
   "odds-ratio",
   "p-value",
   "regression",
   "sample-size",
   "standard-deviation",
   "standard-error",
   "statistical-power",
   "statistical-significance",
   "t-test",
   "variance"
  ],
  "statistic_groups": [
   {
    "g": "central",
    "label": "central tendency",
    "terms": [
     "mean",
     "median"
    ]
   },
   {
    "g": "dispersion",
    "label": "spread",
    "terms": [
     "confidence-interval",
     "standard-deviation",
     "standard-error",
     "variance"
    ]
   },
   {
    "g": "association",
    "label": "association",
    "terms": [
     "R-squared",
     "correlation",
     "effect-size",
     "odds-ratio",
     "regression"
    ]
   },
   {
    "g": "inference",
    "label": "inference / testing",
    "terms": [
     "ANOVA",
     "chi-square",
     "degrees-of-freedom",
     "hypothesis-testing",
     "null-hypothesis",
     "p-value",
     "sample-size",
     "statistical-power",
     "statistical-significance",
     "t-test"
    ]
   },
   {
    "g": "advanced",
    "label": "advanced",
    "terms": [
     "Bayesian",
     "bootstrap",
     "false-discovery-rate"
    ]
   }
  ],
  "claim_types": [
   "critique",
   "existence",
   "mechanism",
   "method",
   "prediction"
  ]
 },
 "presets": [
  {
   "label": "p-value + standard-deviation, across all fields",
   "stats": [
    "p-value",
    "standard-deviation"
   ],
   "mode": "all"
  },
  {
   "label": "hypothesis-testing anywhere",
   "stats": [
    "hypothesis-testing"
   ],
   "mode": "any"
  },
  {
   "label": "Bayesian methods across disciplines",
   "stats": [
    "Bayesian"
   ],
   "mode": "any"
  },
  {
   "label": "regression + correlation together",
   "stats": [
    "regression",
    "correlation"
   ],
   "mode": "all"
  },
  {
   "label": "effect-size + confidence-interval (modern reporting)",
   "stats": [
    "effect-size",
    "confidence-interval"
   ],
   "mode": "all"
  }
 ],
 "records": [
  {
   "id": "einstein1905-electrodynamics",
   "title": "On the Electrodynamics of Moving Bodies",
   "discipline": "physics",
   "year": 1905,
   "authors": "Albert Einstein",
   "url": "papers/einstein1905-electrodynamics.html",
   "key_terms": [],
   "phenomenon": "",
   "claim_type": "",
   "method": "theory",
   "statistics": [],
   "license_spdx": "public-domain",
   "source_url": "https://www.fourmilab.ch/etexts/einstein/specrel/www/",
   "doi": "",
   "full_text_redistributable": false
  },
  {
   "id": "mendel1866-plant-hybridization",
   "title": "Experiments on Plant Hybridization",
   "discipline": "genetics",
   "year": 1866,
   "authors": "Gregor Mendel",
   "url": "papers/mendel1866-plant-hybridization.html",
   "key_terms": [
    "hybrids",
    "dominant",
    "recessive",
    "characters",
    "offspring",
    "generation",
    "plants",
    "ratio",
    "pollen"
   ],
   "phenomenon": "particulate inheritance (dominant/recessive segregation, 3:1 ratio)",
   "claim_type": "existence",
   "method": "experiment",
   "statistics": [
    {
     "term": "mean",
     "evidence": "According, however, to the law of probability, it will always happen, on the average of many cases, that each pollen form A and a will unite ESP FOUNDATIONS SERIES Versuche über Pl",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/mendel1866-plant-hybridization.html",
      "source_body_sha256": "fae53990f7b7a901a988c0b7dbf3dc8498a41f5040a46c28b2d00c78e014a2f5",
      "evidence_sha256": "23a5c43415926969b5811c7c977890bd1a09d91a6f20f2c6566c43d834007cfe",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "LicenseRef-Public-Domain",
   "license_url": "",
   "source_url": "https://www.mendelweb.org/Mendel.html",
   "doi": "",
   "full_text_redistributable": true
  },
  {
   "id": "turing1936-computable-numbers",
   "title": "On Computable Numbers, with an Application to the Entscheidungsproblem",
   "discipline": "mathematics",
   "year": 1936,
   "authors": "Alan M. Turing",
   "url": "papers/turing1936-computable-numbers.html",
   "key_terms": [
    "computable",
    "machine",
    "numbers",
    "tape",
    "symbols",
    "configuration",
    "finite",
    "universal",
    "decimal"
   ],
   "phenomenon": "computability (the Turing machine and the limits of computation)",
   "claim_type": "existence",
   "method": "mathematical-proof",
   "statistics": [],
   "license_spdx": "LicenseRef-Public-Domain",
   "license_url": "",
   "source_url": "https://www.cs.virginia.edu/~robins/Turing_Paper_1936.pdf",
   "doi": "10.1112/plms/s2-42.1.230",
   "full_text_redistributable": false
  },
  {
   "id": "shannon1948-communication",
   "title": "A Mathematical Theory of Communication",
   "discipline": "information-theory",
   "year": 1948,
   "authors": "Claude E. Shannon",
   "url": "papers/shannon1948-communication.html",
   "key_terms": [
    "communication",
    "information",
    "message",
    "channel",
    "signal",
    "noise",
    "bits",
    "logarithmic measure",
    "transmitter",
    "entropy",
    "discrete"
   ],
   "phenomenon": "channel capacity (the fundamental limit of reliable communication)",
   "claim_type": "existence",
   "method": "theory",
   "statistics": [
    {
     "term": "degrees-of-freedom",
     "evidence": "Let x1 ; : : : ; xn ) be the density distribution function for amplitudes x1 ; : : : ; entropy of the ensemble per degree of freedom by xn at n successive sample points.",
     "decision": "rejected",
     "role": "non-statistical-sense",
     "reason": "lexical match does not denote the controlled statistical concept",
     "provenance": {
      "source_path": "papers/shannon1948-communication.html",
      "source_body_sha256": "b0456f450d0365017c170e2e358c69f6d9b96112f0ededebc327e882d6670d34",
      "evidence_sha256": "9ee66feea0859f659695ef6f6ba7c95ac93ac3117742e22420c399fc838269a7",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "mean",
     "evidence": "ly: W log P N If the noise is Gaussian but with a spectrum which is not necessarily flat, N1 is the geometric mean of (cid:16) (cid:17) the noise power over the various frequencies",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/shannon1948-communication.html",
      "source_body_sha256": "b0456f450d0365017c170e2e358c69f6d9b96112f0ededebc327e882d6670d34",
      "evidence_sha256": "bdab97de38e6965d721c03cb6e75d811a6860ee8a6572d93d6cb3f06bdf226e9",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "sample-size",
     "evidence": "Increasing the sample size always sharpens the possible statistical assertions.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/shannon1948-communication.html",
      "source_body_sha256": "b0456f450d0365017c170e2e358c69f6d9b96112f0ededebc327e882d6670d34",
      "evidence_sha256": "20e03f351ce5c91423b2425b9116ba653ea8d417447908725613e284ca9350a1",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "standard-deviation",
     "evidence": "The ensemble ai ; ) = an (cid:0) sin 2W t (cid:25)( 2W t (cid:0) (cid:25)( (cid:0) with the ai normal and independent all with the same standard deviation pN.",
     "decision": "accepted",
     "role": "substantive-mention",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/shannon1948-communication.html",
      "source_body_sha256": "b0456f450d0365017c170e2e358c69f6d9b96112f0ededebc327e882d6670d34",
      "evidence_sha256": "db977f5a3163f4bdd008b9b19f29f5e08515d7ede6016527dca6600a8798972f",
      "grounding_status": "not-exactly-located"
     }
    }
   ],
   "license_spdx": "open-access",
   "license_url": "",
   "source_url": "https://people.math.harvard.edu/~ctm/home/text/others/shannon/entropy/entropy.pdf",
   "doi": "10.1002/j.1538-7305.1948.tb01338.x",
   "full_text_redistributable": false
  },
  {
   "id": "turing1950-machinery-intelligence",
   "title": "Computing Machinery and Intelligence",
   "discipline": "artificial-intelligence",
   "year": 1950,
   "authors": "Alan M. Turing",
   "url": "papers/turing1950-machinery-intelligence.html",
   "key_terms": [
    "machine",
    "think",
    "computer",
    "imitation",
    "interrogator",
    "digital",
    "game",
    "intelligence",
    "learning"
   ],
   "phenomenon": "machine intelligence (the imitation game / Turing test)",
   "claim_type": "method",
   "method": "theory",
   "statistics": [],
   "license_spdx": "LicenseRef-Public-Domain",
   "license_url": "",
   "source_url": "https://www.csee.umbc.edu/courses/471/papers/turing.pdf",
   "doi": "10.1093/mind/LIX.236.433",
   "full_text_redistributable": false
  },
  {
   "id": "nash1950-equilibrium",
   "title": "Equilibrium Points in N-Person Games",
   "discipline": "economics",
   "year": 1950,
   "authors": "John F. Nash",
   "url": "papers/nash1950-equilibrium.html",
   "key_terms": [],
   "phenomenon": "",
   "claim_type": "",
   "method": "mathematical-proof",
   "statistics": [],
   "license_spdx": "open-access",
   "source_url": "https://www.pnas.org/doi/10.1073/pnas.36.1.48",
   "doi": "10.1073/pnas.36.1.48",
   "full_text_redistributable": false
  },
  {
   "id": "quine1951-two-dogmas",
   "title": "Two Dogmas of Empiricism",
   "discipline": "philosophy",
   "year": 1951,
   "authors": "Willard Van Orman Quine",
   "url": "papers/quine1951-two-dogmas.html",
   "key_terms": [
    "analytic",
    "synthetic",
    "dogma",
    "reductionism",
    "empiricism",
    "meaning",
    "synonymy",
    "statement",
    "truths"
   ],
   "phenomenon": "the analytic-synthetic distinction",
   "claim_type": "critique",
   "method": "interpretive",
   "statistics": [],
   "license_spdx": "open-access",
   "license_url": "",
   "source_url": "https://www.ditext.com/quine/quine.html",
   "doi": "10.2307/2181906",
   "full_text_redistributable": false
  },
  {
   "id": "codd1970-relational-model",
   "title": "A Relational Model of Data for Large Shared Data Banks",
   "discipline": "computer-science",
   "year": 1970,
   "authors": "Edgar F. Codd",
   "url": "papers/codd1970-relational-model.html",
   "key_terms": [
    "relational model",
    "data independence",
    "relations",
    "normal form",
    "data bank",
    "redundancy",
    "consistency",
    "derivability",
    "data representation",
    "predicate calculus"
   ],
   "phenomenon": "relational data model (data independence via n-ary relations)",
   "claim_type": "method",
   "method": "theory",
   "statistics": [],
   "license_spdx": "open-access",
   "license_url": "",
   "source_url": "https://www.seas.upenn.edu/~zives/03f/cis550/codd.pdf",
   "doi": "10.1145/362384.362685",
   "full_text_redistributable": false
  },
  {
   "id": "kahneman-tversky1979-prospect",
   "title": "Prospect Theory: An Analysis of Decision under Risk",
   "discipline": "psychology",
   "year": 1979,
   "authors": "Daniel Kahneman, Amos Tversky",
   "url": "papers/kahneman-tversky1979-prospect.html",
   "key_terms": [
    "prospect",
    "utility",
    "value",
    "gains",
    "losses",
    "decision",
    "probabilities",
    "risk",
    "certainty"
   ],
   "phenomenon": "decision under risk (choice among risky prospects)",
   "claim_type": "mechanism",
   "method": "experiment",
   "statistics": [
    {
     "term": "Bayesian",
     "evidence": "[12] : Probability and Profit-A Study of Economic Behavior Along Bayesian Lines.",
     "decision": "rejected",
     "role": "bibliography",
     "reason": "occurrence is confined to a citation/reference rather than the article body",
     "provenance": {
      "source_path": "papers/kahneman-tversky1979-prospect.html",
      "source_body_sha256": "fb9c5b36706408de8cc8922ed93f806fc01cc852e20bb348897d0f8501eb0f3f",
      "evidence_sha256": "9e37a28227da24cfcf243e6ed0d8deae4c3731f53260d5af1d0d60a09780f625",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "sample-size",
     "evidence": "N =72 [83]* [17] The data show that 82 per cent of the subjects chose B in Problem 1, and 83 per cent of the subjects chose C in Problem 2.",
     "decision": "rejected",
     "role": "non-statistical-sense",
     "reason": "lexical match does not denote the controlled statistical concept",
     "provenance": {
      "source_path": "papers/kahneman-tversky1979-prospect.html",
      "source_body_sha256": "fb9c5b36706408de8cc8922ed93f806fc01cc852e20bb348897d0f8501eb0f3f",
      "evidence_sha256": "490a64c9b00a31584c4614fbe94a37f3924f8bb179e2af5842fed2c2143d4176",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "statistical-significance",
     "evidence": "Each of these preferences is significant at the .01 level, as denoted by the asterisk.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/kahneman-tversky1979-prospect.html",
      "source_body_sha256": "fb9c5b36706408de8cc8922ed93f806fc01cc852e20bb348897d0f8501eb0f3f",
      "evidence_sha256": "45fc0ffbf5378a68726c629d6777f47e304c7d3aeff038141454ef28c566a050",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "variance",
     "evidence": "When the prospects are reduced, however, the difference in variance between (3,000,.25) and (4,000,.20) may be insufficient to overcome the difference in expected value.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/kahneman-tversky1979-prospect.html",
      "source_body_sha256": "fb9c5b36706408de8cc8922ed93f806fc01cc852e20bb348897d0f8501eb0f3f",
      "evidence_sha256": "f082f4d74cc88ded6a3d68d533245f8af604f950131731ef42bb544fc7cc0dd0",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "open-access",
   "license_url": "",
   "source_url": "https://www.uzh.ch/cmsssl/suz/dam/jcr:00000000-64a0-5b1c-0000-00003b7ec704/10.05-kahneman-tversky-79.pdf",
   "doi": "10.2307/1914185",
   "full_text_redistributable": false
  },
  {
   "id": "goodfellow2014-gan",
   "title": "Generative Adversarial Networks",
   "discipline": "artificial-intelligence",
   "year": 2014,
   "authors": "Ian J. Goodfellow, Jean Pouget-Abadie, Mehdi Mirza, Bing Xu, David Warde-Farley, Sherjil Ozair, Aaron Courville, Yoshua Bengio",
   "url": "papers/goodfellow2014-gan.html",
   "key_terms": [
    "generative",
    "adversarial",
    "discriminator",
    "generator",
    "distribution",
    "minimax",
    "probability",
    "training",
    "samples"
   ],
   "phenomenon": "generative adversarial networks (adversarial training; minimax game)",
   "claim_type": "method",
   "method": "computational-model",
   "statistics": [
    {
     "term": "Bayesian",
     "evidence": "More recent examples of training a generative machine by back-propagating into it include recent work on auto-encoding variational Bayes [20] and stochastic backpropagation [24].",
     "decision": "accepted",
     "role": "substantive-mention",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/goodfellow2014-gan.html",
      "source_body_sha256": "043e6e10fe1edeb5abc75c705df0738e830524b4e80f26de932162f8dda8cae9",
      "evidence_sha256": "486c3d5cb11357b3c38490415b5dc6141e2648ae5d63b9f53e4acc93840cda37",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "standard-error",
     "evidence": "The reported numbers on MNIST are the mean log- likelihood of samples on test set, with the standard error of the mean computed across examples.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit standard error or SEM",
     "provenance": {
      "source_path": "papers/goodfellow2014-gan.html",
      "source_body_sha256": "043e6e10fe1edeb5abc75c705df0738e830524b4e80f26de932162f8dda8cae9",
      "evidence_sha256": "eec3d63868a2cfa632f14e49672ea34f6fd67f6d70ed5acb3bc3c8ededcb12b4",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "variance",
     "evidence": "This method of estimating the likelihood has somewhat high variance and does not perform well in high dimensional spaces but it is the best method available to our knowledge.",
     "decision": "accepted",
     "role": "substantive-mention",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/goodfellow2014-gan.html",
      "source_body_sha256": "043e6e10fe1edeb5abc75c705df0738e830524b4e80f26de932162f8dda8cae9",
      "evidence_sha256": "4f89583375e52e88e3abfd6efe50d3e18707521229559c9d56b87c696e09754b",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "arXiv-nonexclusive-distrib-1.0",
   "license_url": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
   "source_url": "https://arxiv.org/abs/1406.2661",
   "doi": "",
   "full_text_redistributable": false
  },
  {
   "id": "vaswani2017-attention",
   "title": "Attention Is All You Need",
   "discipline": "artificial-intelligence",
   "year": 2017,
   "authors": "Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N. Gomez, Lukasz Kaiser, Illia Polosukhin",
   "url": "papers/vaswani2017-attention.html",
   "key_terms": [
    "attention",
    "Transformer",
    "encoder",
    "decoder",
    "recurrence",
    "convolutions",
    "sequence",
    "translation",
    "parallelizable"
   ],
   "phenomenon": "sequence transduction with attention (the Transformer; attention-only architecture)",
   "claim_type": "method",
   "method": "computational-model",
   "statistics": [
    {
     "term": "statistical-significance",
     "evidence": "Experiments on two machine translation tasks show these models to be superior in quality while being more parallelizable and requiring significantly less time to train.",
     "decision": "rejected",
     "role": "lexical-significance",
     "reason": "significant/significantly is used without inferential-statistical meaning",
     "provenance": {
      "source_path": "papers/vaswani2017-attention.html",
      "source_body_sha256": "5f8f0f159e841ca84efdcf90ba8e6f98d8248ee6540e55ab1297cb9e1c921f09",
      "evidence_sha256": "5493cb66eeb564a5799343f12fb7a8f78254372b6ded0dd613519f3ea1d34ce9",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "variance",
     "evidence": "dot products get large, assume that the components of q and k are independent random i=1 qiki, has mean 0 and variance dk.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/vaswani2017-attention.html",
      "source_body_sha256": "5f8f0f159e841ca84efdcf90ba8e6f98d8248ee6540e55ab1297cb9e1c921f09",
      "evidence_sha256": "2118da995bd1135e502ed3b960bf2ead4416e01ecfb17d5a958769cb5bf6ccc0",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "arXiv-nonexclusive-distrib-1.0",
   "license_url": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
   "source_url": "https://arxiv.org/abs/1706.03762",
   "doi": "",
   "full_text_redistributable": false
  },
  {
   "id": "brown2020-gpt3",
   "title": "Language Models are Few-Shot Learners",
   "discipline": "artificial-intelligence",
   "year": 2020,
   "authors": "Tom B. Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah, et al. (OpenAI)",
   "url": "papers/brown2020-gpt3.html",
   "key_terms": [
    "language models",
    "few-shot",
    "in-context learning",
    "fine-tuning",
    "meta-learning",
    "zero-shot",
    "one-shot",
    "autoregressive",
    "pre-training",
    "Common Crawl",
    "task-agnostic",
    "demonstrations"
   ],
   "phenomenon": "in-context / few-shot learning in large language models (emergent from scale)",
   "claim_type": "existence",
   "method": "computational-model",
   "statistics": [
    {
     "term": "confidence-interval",
     "evidence": "25 Mean accuracy 95% Confidence Interval (low, hi) t compared to control (p-value) “I don’t know” assignments Control (deliberately bad model) GPT-3 Small GPT-3 Medium GPT-3 Large ",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/brown2020-gpt3.html",
      "source_body_sha256": "b378cbecbfd49cb0c222b262a4902162c36ae9ccdbc3ab02d4084a69ad98ee81",
      "evidence_sha256": "cba20de7325fd3c0ab688074ea26a8a0cc1b17ff24c47bd58ab19258a519fe23",
      "grounding_status": "not-exactly-located"
     }
    },
    {
     "term": "effect-size",
     "evidence": "• German translation: We found 25% of the examples in the WMT16 German-English test set were marked as potentially contaminated, with an associated total effect size of 1-2 BLEU.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/brown2020-gpt3.html",
      "source_body_sha256": "b378cbecbfd49cb0c222b262a4902162c36ae9ccdbc3ab02d4084a69ad98ee81",
      "evidence_sha256": "0bf74b14033314da0f10d169e789ab6db49079ffc9da1a5ac2c53bfc8b25000b",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "hypothesis-testing",
     "evidence": "In this paper, we test this hypothesis by training a 175 billion parameter autoregressive language model, which we call GPT-3, and measuring its in-context learning abilities.",
     "decision": "rejected",
     "role": "non-statistical-sense",
     "reason": "lexical match does not denote the controlled statistical concept",
     "provenance": {
      "source_path": "papers/brown2020-gpt3.html",
      "source_body_sha256": "b378cbecbfd49cb0c222b262a4902162c36ae9ccdbc3ab02d4084a69ad98ee81",
      "evidence_sha256": "adb14b4f82dc0a9a95ad00a8dfbf53ec02edf7e3ac93d662cf32ff1965def7ed",
      "grounding_status": "not-exactly-located"
     }
    },
    {
     "term": "mean",
     "evidence": "The 95% confidence intervals were estimated from the t-distribution of the sample mean.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/brown2020-gpt3.html",
      "source_body_sha256": "b378cbecbfd49cb0c222b262a4902162c36ae9ccdbc3ab02d4084a69ad98ee81",
      "evidence_sha256": "6cc1750e72a7207f4ff81bd81d742acb04ebf6a48b50e563fdc26dde2b25cad9",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "p-value",
     "evidence": "25 Mean accuracy 95% Confidence Interval (low, hi) t compared to control (p-value) “I don’t know” assignments Control (deliberately bad model) GPT-3 Small GPT-3 Medium GPT-3 Large ",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit p-value or inferential p notation",
     "provenance": {
      "source_path": "papers/brown2020-gpt3.html",
      "source_body_sha256": "b378cbecbfd49cb0c222b262a4902162c36ae9ccdbc3ab02d4084a69ad98ee81",
      "evidence_sha256": "cba20de7325fd3c0ab688074ea26a8a0cc1b17ff24c47bd58ab19258a519fe23",
      "grounding_status": "not-exactly-located"
     }
    },
    {
     "term": "regression",
     "evidence": "The classifier is trained using logistic regression classifier with features from Spark’s standard tokenizer and HashingTF 10.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/brown2020-gpt3.html",
      "source_body_sha256": "b378cbecbfd49cb0c222b262a4902162c36ae9ccdbc3ab02d4084a69ad98ee81",
      "evidence_sha256": "c2b6fc79b0a37f7bbacaed0f07c2b079b5a1306449fe492bd6a83191dd9950f0",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "standard-deviation",
     "evidence": "Results are on the dev-set, which has only 1500 examples and therefore has high variance (we estimate a standard deviation of 1.2%).",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/brown2020-gpt3.html",
      "source_body_sha256": "b378cbecbfd49cb0c222b262a4902162c36ae9ccdbc3ab02d4084a69ad98ee81",
      "evidence_sha256": "e1dd769d070906e36a02e0c5b91ccec5767bb45b197ac0371d84fd9d673d0cb5",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "statistical-significance",
     "evidence": "We provide one to five previous examples of a (separate) 5We use a two-sample Student’s T-Test to test for significant difference between the means of the participant accuracies of",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit inferential-statistical context",
     "provenance": {
      "source_path": "papers/brown2020-gpt3.html",
      "source_body_sha256": "b378cbecbfd49cb0c222b262a4902162c36ae9ccdbc3ab02d4084a69ad98ee81",
      "evidence_sha256": "7b8a87606e3a32252b866ccf9ab4c5fa17bbe4bde5e9a002012fe645a3d00bc2",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "t-test",
     "evidence": "This table compares mean accuracy between five different models, and shows the results of a two-sample T-Test for the difference in mean accuracy between each model and the control",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/brown2020-gpt3.html",
      "source_body_sha256": "b378cbecbfd49cb0c222b262a4902162c36ae9ccdbc3ab02d4084a69ad98ee81",
      "evidence_sha256": "80dc85d3b61269423d270e4042666d2634531032263b402c8e77cb83c0774070",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "variance",
     "evidence": "Results are on the dev-set, which has only 1500 examples and therefore has high variance (we estimate a standard deviation of 1.2%).",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/brown2020-gpt3.html",
      "source_body_sha256": "b378cbecbfd49cb0c222b262a4902162c36ae9ccdbc3ab02d4084a69ad98ee81",
      "evidence_sha256": "e1dd769d070906e36a02e0c5b91ccec5767bb45b197ac0371d84fd9d673d0cb5",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "arXiv-nonexclusive-distrib-1.0",
   "license_url": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
   "source_url": "https://arxiv.org/abs/2005.14165",
   "doi": "",
   "full_text_redistributable": false
  },
  {
   "id": "bai2022-constitutional-ai",
   "title": "Constitutional AI: Harmlessness from AI Feedback",
   "discipline": "artificial-intelligence",
   "year": 2022,
   "authors": "Yuntao Bai, Saurav Kadavath, Sandipan Kundu, et al. (Anthropic)",
   "url": "papers/bai2022-constitutional-ai.html",
   "key_terms": [
    "Constitutional AI",
    "harmlessness",
    "reinforcement learning",
    "AI feedback",
    "preference model",
    "supervised learning",
    "self-critique",
    "RLAIF",
    "chain-of-thought",
    "scaling supervision",
    "principles"
   ],
   "phenomenon": "AI alignment / harmlessness (Constitutional AI; training with principles and AI feedback)",
   "claim_type": "method",
   "method": "computational-model",
   "statistics": [
    {
     "term": "statistical-significance",
     "evidence": "Past experiments [Bai et al., 2022] showed that RLHF significantly improves the models’ ability to follow instructions, and the HH model is significantly more harmless than the hel",
     "decision": "rejected",
     "role": "lexical-significance",
     "reason": "significant/significantly is used without inferential-statistical meaning",
     "provenance": {
      "source_path": "papers/bai2022-constitutional-ai.html",
      "source_body_sha256": "25ebfa88fda565c2651a961552e77a22cd65454ef7d4d02eec1554717034b41a",
      "evidence_sha256": "0e22eeb47740f95b1e7a9c87e0c63981dbcb3168f1b94098700e628ca3532496",
      "grounding_status": "not-exactly-located"
     }
    }
   ],
   "license_spdx": "CC-BY-4.0",
   "license_url": "http://creativecommons.org/licenses/by/4.0/",
   "source_url": "https://arxiv.org/abs/2212.08073",
   "doi": "",
   "full_text_redistributable": true
  },
  {
   "id": "barabasi-albert1999-scale-free",
   "title": "Emergence of Scaling in Random Networks",
   "discipline": "sociology",
   "year": 1999,
   "authors": "Albert-László Barabási, Réka Albert",
   "url": "papers/barabasi-albert1999-scale-free.html",
   "key_terms": [
    "network",
    "vertices",
    "power-law",
    "scale-free",
    "preferential",
    "exponent",
    "edges",
    "distribution",
    "probability"
   ],
   "phenomenon": "scale-free network (power-law degree distribution)",
   "claim_type": "mechanism",
   "method": "computational-model",
   "statistics": [],
   "license_spdx": "arxiv-open",
   "license_url": "",
   "source_url": "https://arxiv.org/abs/cond-mat/9910332",
   "doi": "10.1126/science.286.5439.509",
   "full_text_redistributable": false
  },
  {
   "id": "cook2013-climate-consensus",
   "title": "Quantifying the Consensus on Anthropogenic Global Warming in the Scientific Literature",
   "discipline": "earth-science",
   "year": 2013,
   "authors": "John Cook, Dana Nuccitelli, Sarah A. Green, et al.",
   "url": "papers/cook2013-climate-consensus.html",
   "key_terms": [
    "consensus",
    "abstracts",
    "AGW",
    "warming",
    "climate",
    "endorsed",
    "papers",
    "position",
    "anthropogenic"
   ],
   "phenomenon": "scientific consensus on anthropogenic global warming (AGW)",
   "claim_type": "existence",
   "method": "corpus-analysis",
   "statistics": [
    {
     "term": "confidence-interval",
     "evidence": "percentage has increased (simple linear regression trend 0.87% ± 0.28% yr1, 95% CI, R2 = 0.66, p < 0.001) and the percentage of papers taking a position on AGW has equally decrease",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/cook2013-climate-consensus.html",
      "source_body_sha256": "1ab6733b115f76263fbdc9f91a4e18662d12ea83a1f168daf5292b834034abc2",
      "evidence_sha256": "a861714cfa8f06e30e3a532b68cd0727e145976470049233fa251a78eaaadc65",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "p-value",
     "evidence": "percentage has increased (simple linear regression trend 0.87% ± 0.28% yr1, 95% CI, R2 = 0.66, p < 0.001) and the percentage of papers taking a position on AGW has equally decrease",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit p-value or inferential p notation",
     "provenance": {
      "source_path": "papers/cook2013-climate-consensus.html",
      "source_body_sha256": "1ab6733b115f76263fbdc9f91a4e18662d12ea83a1f168daf5292b834034abc2",
      "evidence_sha256": "a861714cfa8f06e30e3a532b68cd0727e145976470049233fa251a78eaaadc65",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "regression",
     "evidence": "percentage has increased (simple linear regression trend 0.87% ± 0.28% yr1, 95% CI, R2 = 0.66, p < 0.001) and the percentage of papers taking a position on AGW has equally decrease",
     "decision": "accepted",
     "role": "substantive-mention",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/cook2013-climate-consensus.html",
      "source_body_sha256": "1ab6733b115f76263fbdc9f91a4e18662d12ea83a1f168daf5292b834034abc2",
      "evidence_sha256": "a861714cfa8f06e30e3a532b68cd0727e145976470049233fa251a78eaaadc65",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "sample-size",
     "evidence": "The year-to-year endorsing or variability is larger in the self-ratings than in the abstract ratings due to the smaller sample sizes in the early 1990s.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/cook2013-climate-consensus.html",
      "source_body_sha256": "1ab6733b115f76263fbdc9f91a4e18662d12ea83a1f168daf5292b834034abc2",
      "evidence_sha256": "98dbf2f8df76379e3eb69c1585f8ba0b03ccb001575769a8910c7936f9db4e4f",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "statistical-significance",
     "evidence": "The average numbers of authors per endorsement abstract (3.4) and per no position abstract (3.6) are both significantly larger than the average number of authors per rejection abst",
     "decision": "rejected",
     "role": "lexical-significance",
     "reason": "significant/significantly is used without inferential-statistical meaning",
     "provenance": {
      "source_path": "papers/cook2013-climate-consensus.html",
      "source_body_sha256": "1ab6733b115f76263fbdc9f91a4e18662d12ea83a1f168daf5292b834034abc2",
      "evidence_sha256": "c414227405f0b05561a1e3c757021b2b75156585f4767dea65fdfa5a2574f8da",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "CC-BY-3.0",
   "license_url": "http://creativecommons.org/licenses/by/3.0",
   "source_url": "https://iopscience.iop.org/article/10.1088/1748-9326/8/2/024024",
   "doi": "10.1088/1748-9326/8/2/024024",
   "full_text_redistributable": true
  },
  {
   "id": "ioannidis2005-false-findings",
   "title": "Why Most Published Research Findings Are False",
   "discipline": "statistics",
   "year": 2005,
   "authors": "John P. A. Ioannidis",
   "url": "papers/ioannidis2005-false-findings.html",
   "key_terms": [
    "research",
    "studies",
    "power",
    "bias",
    "probability",
    "true",
    "false",
    "relationships",
    "PPV"
   ],
   "phenomenon": "false / irreproducible published research findings",
   "claim_type": "mechanism",
   "method": "formal-model",
   "statistics": [
    {
     "term": "Bayesian",
     "evidence": "As has been shown previously, the probability that a research fi nding is indeed true depends on the prior probability of it being true (before doing the study), the statistical po",
     "decision": "rejected",
     "role": "insufficient-evidence",
     "reason": "supplied evidence does not establish the controlled term in the article body",
     "provenance": {
      "source_path": "papers/ioannidis2005-false-findings.html",
      "source_body_sha256": "4f5f35b2ebc644587beb285cb7bb851fc5bd5fdbc39b437e8f6dbbfc3c682ae4",
      "evidence_sha256": "1a17ee9177e86fd8d530fb1756923f8d564bc7630130d7eff5d42025c2b26382",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "effect-size",
     "evidence": "ework, a research fi nding is less likely to be true when the studies conducted in a fi eld are smaller; when effect sizes are smaller; when there is a greater number and lesser pr",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/ioannidis2005-false-findings.html",
      "source_body_sha256": "4f5f35b2ebc644587beb285cb7bb851fc5bd5fdbc39b437e8f6dbbfc3c682ae4",
      "evidence_sha256": "01a2fc14723065f038eba8a509acd8347671c2451eb90e6d16620e6f9f57d9f9",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "false-discovery-rate",
     "evidence": "Despite a large statistical literature for multiple testing corrections [37], usually it is impossible to decipher how much data dredging by the reporting authors or other research",
     "decision": "rejected",
     "role": "non-statistical-sense",
     "reason": "lexical match does not denote the controlled statistical concept",
     "provenance": {
      "source_path": "papers/ioannidis2005-false-findings.html",
      "source_body_sha256": "4f5f35b2ebc644587beb285cb7bb851fc5bd5fdbc39b437e8f6dbbfc3c682ae4",
      "evidence_sha256": "abb0139ad6a36d68ed071407c74b1774a738b860acb25caf646144f2103fdf7b",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "hypothesis-testing",
     "evidence": "Hsueh HM, Chen JJ, Kodell RL (2003) Comparison of methods for estimating the number of true null hypotheses in multiplicity testing.",
     "decision": "rejected",
     "role": "bibliography",
     "reason": "occurrence is confined to a citation/reference rather than the article body",
     "provenance": {
      "source_path": "papers/ioannidis2005-false-findings.html",
      "source_body_sha256": "4f5f35b2ebc644587beb285cb7bb851fc5bd5fdbc39b437e8f6dbbfc3c682ae4",
      "evidence_sha256": "ddd6a99face09234ac16baa5098a9c432a1b5a3a0df626bd4c39004efcde57e8",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "null-hypothesis",
     "evidence": "Hsueh HM, Chen JJ, Kodell RL (2003) Comparison of methods for estimating the number of true null hypotheses in multiplicity testing.",
     "decision": "rejected",
     "role": "bibliography",
     "reason": "occurrence is confined to a citation/reference rather than the article body",
     "provenance": {
      "source_path": "papers/ioannidis2005-false-findings.html",
      "source_body_sha256": "4f5f35b2ebc644587beb285cb7bb851fc5bd5fdbc39b437e8f6dbbfc3c682ae4",
      "evidence_sha256": "ddd6a99face09234ac16baa5098a9c432a1b5a3a0df626bd4c39004efcde57e8",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "odds-ratio",
     "evidence": "n gene polymorphisms among those tested would be truly associated with schizophrenia, with relatively similar odds ratios around 1.3 for the ten or so polymorphisms and with a fair",
     "decision": "accepted",
     "role": "substantive-mention",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/ioannidis2005-false-findings.html",
      "source_body_sha256": "4f5f35b2ebc644587beb285cb7bb851fc5bd5fdbc39b437e8f6dbbfc3c682ae4",
      "evidence_sha256": "7bc507787ca12e40ac73d14d3cd4f06c9f8a3b93aea16fe59d652a01e2bb5f17",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "p-value",
     "evidence": "-founded strategy of claiming conclusive research fi ndings solely on the basis of a single study assessed by formal statistical signifi cance, typically for a p-value less than 0.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit p-value or inferential p notation",
     "provenance": {
      "source_path": "papers/ioannidis2005-false-findings.html",
      "source_body_sha256": "4f5f35b2ebc644587beb285cb7bb851fc5bd5fdbc39b437e8f6dbbfc3c682ae4",
      "evidence_sha256": "24ec8d9a20812b4dc105a9c760da12bc94610fb565b7f002c34433f103858092",
      "grounding_status": "not-exactly-located"
     }
    },
    {
     "term": "sample-size",
     "evidence": "Small sample size means smaller power and, for all functions above, the PPV for a true research fi nding decreases as power decreases towards 1 − β = 0.05.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/ioannidis2005-false-findings.html",
      "source_body_sha256": "4f5f35b2ebc644587beb285cb7bb851fc5bd5fdbc39b437e8f6dbbfc3c682ae4",
      "evidence_sha256": "e8cbcb5dd559f0f9fa24135378a37e75607723fcccb703343cbd363632f82a98",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "statistical-power",
     "evidence": "earch fi nding is indeed true depends on the prior probability of it being true (before doing the study), the statistical power of the study, and the level of statistical signifi c",
     "decision": "accepted",
     "role": "substantive-mention",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/ioannidis2005-false-findings.html",
      "source_body_sha256": "4f5f35b2ebc644587beb285cb7bb851fc5bd5fdbc39b437e8f6dbbfc3c682ae4",
      "evidence_sha256": "7b4174e8749eed4063fa969ac039298f1ad00ec213a09ab6d2db163b7f051a47",
      "grounding_status": "not-exactly-located"
     }
    }
   ],
   "license_spdx": "CC-BY-4.0",
   "license_url": "http://creativecommons.org/licenses/by/4.0/",
   "source_url": "https://journals.plos.org/plosmedicine/article?id=10.1371/journal.pmed.0020124",
   "doi": "10.1371/journal.pmed.0020124",
   "full_text_redistributable": true
  },
  {
   "id": "sporns2005-human-connectome",
   "title": "The Human Connectome: A Structural Description of the Human Brain",
   "discipline": "neuroscience",
   "year": 2005,
   "authors": "Olaf Sporns, Giulio Tononi, Rolf Kötter",
   "url": "papers/sporns2005-human-connectome.html",
   "key_terms": [
    "connectome",
    "brain",
    "network",
    "connections",
    "structural",
    "neurons",
    "connectivity",
    "nodes",
    "matrix"
   ],
   "phenomenon": "brain structural connectivity network (the connectome)",
   "claim_type": "method",
   "method": "theory",
   "statistics": [],
   "license_spdx": "CC-BY-4.0",
   "license_url": "http://creativecommons.org/licenses/by/4.0/",
   "source_url": "https://journals.plos.org/ploscompbiol/article?id=10.1371/journal.pcbi.0010042",
   "doi": "10.1371/journal.pcbi.0010042",
   "full_text_redistributable": true
  },
  {
   "id": "gomez-bombarelli2018-molecular-design",
   "title": "Automatic Chemical Design Using a Data-Driven Continuous Representation of Molecules",
   "discipline": "chemistry",
   "year": 2018,
   "authors": "Rafael Gómez-Bombarelli, Jennifer N. Wei, David Duvenaud, et al.",
   "url": "papers/gomez-bombarelli2018-molecular-design.html",
   "key_terms": [
    "molecules",
    "continuous",
    "representation",
    "encoder",
    "decoder",
    "chemical",
    "latent",
    "optimization",
    "SMILES"
   ],
   "phenomenon": "generative molecular design (continuous latent representation of molecules; VAE)",
   "claim_type": "method",
   "method": "computational-model",
   "statistics": [
    {
     "term": "Bayesian",
     "evidence": "Gradient-based optimization can be combined with Bayesian optimization methods to select compounds that are likely to be informative about the global optimum.",
     "decision": "accepted",
     "role": "substantive-mention",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/gomez-bombarelli2018-molecular-design.html",
      "source_body_sha256": "afa86132cd428d0ca4ff398a172ea554983ec8212dfbe23fba26f40e03362b48",
      "evidence_sha256": "bce0397fec5a1602dc1623c44b47e05b1fe0ebd59e4c9c7ce459e9ec0aa561eb",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "mean",
     "evidence": "In addition to the added noise to the encoder, which can be turned off by simply sampling the mean of the encoding distribution, the decoder also samples a string from of the proba",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "numeric mean or average in main text",
     "provenance": {
      "source_path": "papers/gomez-bombarelli2018-molecular-design.html",
      "source_body_sha256": "afa86132cd428d0ca4ff398a172ea554983ec8212dfbe23fba26f40e03362b48",
      "evidence_sha256": "4ce85568deefce19ca1edec2eb9a83b734e059517fb56bcaae59c2108e8ff8af",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "regression",
     "evidence": "omatically build an even larger implicit library, and then use the smaller set of labeled examples to build a regression model from the continuous representation to the desired pro",
     "decision": "accepted",
     "role": "substantive-mention",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/gomez-bombarelli2018-molecular-design.html",
      "source_body_sha256": "afa86132cd428d0ca4ff398a172ea554983ec8212dfbe23fba26f40e03362b48",
      "evidence_sha256": "6a85ca027766d355cb39464a0ca4548a5fd9e827b241989d2f48357f55a22312",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "standard-deviation",
     "evidence": "Whereas each individual dimension shows a slightly different mean and standard deviation, all follow normal distribution as enforced by the variational regularizer.",
     "decision": "accepted",
     "role": "substantive-mention",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/gomez-bombarelli2018-molecular-design.html",
      "source_body_sha256": "afa86132cd428d0ca4ff398a172ea554983ec8212dfbe23fba26f40e03362b48",
      "evidence_sha256": "4c322ee2280b3a87a9bea4f41d19938ecd5586bc3619948f2694b0cce0493b6e",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "arXiv-nonexclusive-distrib-1.0",
   "license_url": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
   "source_url": "https://arxiv.org/abs/1610.02415",
   "doi": "10.1021/acscentsci.7b00572",
   "full_text_redistributable": false
  },
  {
   "id": "schmidhuber2015-deep-learning",
   "title": "Deep Learning in Neural Networks: An Overview",
   "discipline": "artificial-intelligence",
   "year": 2015,
   "authors": "Jürgen Schmidhuber",
   "url": "papers/schmidhuber2015-deep-learning.html",
   "key_terms": [
    "deep learning",
    "neural networks",
    "credit assignment",
    "supervised learning",
    "unsupervised learning",
    "reinforcement learning",
    "backpropagation",
    "pattern recognition",
    "recurrent",
    "depth"
   ],
   "phenomenon": "deep learning (credit assignment path depth)",
   "claim_type": "method",
   "method": "theory",
   "statistics": [
    {
     "term": "Bayesian",
     "evidence": "NN case, the MDL princi- ple suggests that low NN weight complexity corresponds to high NN probability in the Bayesian view (e.g., MacKay, 1992; Buntine and Weigend, 1991; Neal, 19",
     "decision": "accepted",
     "role": "substantive-mention",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/schmidhuber2015-deep-learning.html",
      "source_body_sha256": "f62bc15adb4d3ded421a6e6a63b764df16f07d4707238f309059b4cdc48fb938",
      "evidence_sha256": "d79491ae13df00ccbb9c949f78dc0498445b47b3873a64f1e1b281b6b47fefec",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "mean",
     "evidence": "Lecture 6.5—RmsProp: Divide the gradient by a running average of its recent magnitude.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "numeric mean or average in main text",
     "provenance": {
      "source_path": "papers/schmidhuber2015-deep-learning.html",
      "source_body_sha256": "f62bc15adb4d3ded421a6e6a63b764df16f07d4707238f309059b4cdc48fb938",
      "evidence_sha256": "e22b7174f4692eee64a5ac5ab994cdfd4ec2c4805552379b96a96fc567bca65b",
      "grounding_status": "verified-reference"
     }
    },
    {
     "term": "regression",
     "evidence": "In a sense NNs have been around even longer, since early supervised NNs were essentially variants of linear regression methods going back at least to the early 1800s (e.g., Legendr",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/schmidhuber2015-deep-learning.html",
      "source_body_sha256": "f62bc15adb4d3ded421a6e6a63b764df16f07d4707238f309059b4cdc48fb938",
      "evidence_sha256": "11363456bef3892dbe8f097077ffd898e8c2447cc3e0b77bf239e1df68d034d9",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "variance",
     "evidence": "Few bits of information are required to describe low-precision weights with high variance.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/schmidhuber2015-deep-learning.html",
      "source_body_sha256": "f62bc15adb4d3ded421a6e6a63b764df16f07d4707238f309059b4cdc48fb938",
      "evidence_sha256": "dccc5ca8e18db10d7bd95cd8e860df5fa29b6065c4eab2113668409498aaf318",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "arXiv-nonexclusive-distrib-1.0",
   "license_url": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
   "source_url": "https://arxiv.org/abs/1404.7828",
   "doi": "10.1016/j.neunet.2014.09.003",
   "full_text_redistributable": false
  },
  {
   "id": "stephens2015-genomical",
   "title": "Big Data: Astronomical or Genomical?",
   "discipline": "biology",
   "year": 2015,
   "authors": "Zachary D. Stephens, Skylar Y. Lee, Faraz Faghri, et al.",
   "url": "papers/stephens2015-genomical.html",
   "key_terms": [
    "genomics",
    "Big Data",
    "sequencing",
    "data acquisition",
    "astronomy",
    "genome",
    "storage",
    "petabases",
    "exabase",
    "computational challenges"
   ],
   "phenomenon": "Big Data growth (exponential data generation across domains)",
   "claim_type": "prediction",
   "method": "theory",
   "statistics": [],
   "license_spdx": "CC-BY-4.0",
   "license_url": "http://creativecommons.org/licenses/by/4.0/",
   "source_url": "https://journals.plos.org/plosbiology/article?id=10.1371/journal.pbio.1002195",
   "doi": "10.1371/journal.pbio.1002195",
   "full_text_redistributable": true
  },
  {
   "id": "mikolov2013-word2vec",
   "title": "Efficient Estimation of Word Representations in Vector Space",
   "discipline": "linguistics",
   "year": 2013,
   "authors": "Tomas Mikolov, Kai Chen, Greg Corrado, Jeffrey Dean",
   "url": "papers/mikolov2013-word2vec.html",
   "key_terms": [
    "word",
    "vector",
    "representations",
    "semantic",
    "syntactic",
    "similarity",
    "architecture",
    "CBOW",
    "Skip-gram"
   ],
   "phenomenon": "word embeddings (distributed vector representations of words)",
   "claim_type": "method",
   "method": "computational-model",
   "statistics": [
    {
     "term": "correlation",
     "evidence": "The publicly available RNN vectors were used together with other techniques to achieve over 50% increase in Spearman’s rank correlation over the previous best result [31].",
     "decision": "accepted",
     "role": "substantive-mention",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/mikolov2013-word2vec.html",
      "source_body_sha256": "65551602a57879c73147f11e0cf269749e158d39fa1191582816019343e8182f",
      "evidence_sha256": "cca392c288b153669b2d99929879a1a3c86b1a58c6b0ca5c28230f3a858e9493",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "statistical-significance",
     "evidence": "However, as far as we know, these architectures were significantly more computationally expensive for training than the one proposed in [13], with the exception of certain vers",
     "decision": "rejected",
     "role": "lexical-significance",
     "reason": "significant/significantly is used without inferential-statistical meaning",
     "provenance": {
      "source_path": "papers/mikolov2013-word2vec.html",
      "source_body_sha256": "65551602a57879c73147f11e0cf269749e158d39fa1191582816019343e8182f",
      "evidence_sha256": "6a60108911386f4ff7a206c14eb7b9218567f6f5050b28ca95209e8456bede19",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "arXiv-nonexclusive-distrib-1.0",
   "license_url": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
   "source_url": "https://arxiv.org/abs/1301.3781",
   "doi": "",
   "full_text_redistributable": false
  },
  {
   "id": "kleinberg2016-fairness",
   "title": "Inherent Trade-Offs in the Fair Determination of Risk Scores",
   "discipline": "law",
   "year": 2016,
   "authors": "Jon Kleinberg, Sendhil Mullainathan, Manish Raghavan",
   "url": "papers/kleinberg2016-fairness.html",
   "key_terms": [
    "fairness",
    "groups",
    "calibrated",
    "risk",
    "conditions",
    "classification",
    "recidivism",
    "defendants"
   ],
   "phenomenon": "algorithmic fairness (inherent trade-offs between fairness criteria)",
   "claim_type": "existence",
   "method": "mathematical-proof",
   "statistics": [
    {
     "term": "Bayesian",
     "evidence": "Three naive bayes approaches for discrimination-free classification.",
     "decision": "rejected",
     "role": "bibliography",
     "reason": "occurrence is confined to a citation/reference rather than the article body",
     "provenance": {
      "source_path": "papers/kleinberg2016-fairness.html",
      "source_body_sha256": "c457861474d0e12765a8aee35745a9edd35d1d7124db3865d18d60476c25d485",
      "evidence_sha256": "36b56412997bccdd411feba4b7a882ce627791fb13bcf9c47bc164ce7b47d1cd",
      "grounding_status": "verified-reference"
     }
    },
    {
     "term": "mean",
     "evidence": "Suppose, alternately, that the two groups have the same fraction of members in the positive class; that is, the average value of pσ is the same for the members of group 1 and group",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/kleinberg2016-fairness.html",
      "source_body_sha256": "c457861474d0e12765a8aee35745a9edd35d1d7124db3865d18d60476c25d485",
      "evidence_sha256": "820f434c3a8ed27c2f5228887a1f77aa8050b429161ae4a3a647511ef19a1a43",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "statistical-significance",
     "evidence": "tings where these algorithmic and statistical techniques, trained on large datasets of past instances, play a significant role in the outcome.",
     "decision": "rejected",
     "role": "lexical-significance",
     "reason": "significant/significantly is used without inferential-statistical meaning",
     "provenance": {
      "source_path": "papers/kleinberg2016-fairness.html",
      "source_body_sha256": "c457861474d0e12765a8aee35745a9edd35d1d7124db3865d18d60476c25d485",
      "evidence_sha256": "8f6d1bd3956cfecb094d455514c75de75c371629e9234c8118e51ce5d9a70307",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "arXiv-nonexclusive-distrib-1.0",
   "license_url": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
   "source_url": "https://arxiv.org/abs/1609.05807",
   "doi": "",
   "full_text_redistributable": false
  },
  {
   "id": "amodei2016-ai-safety",
   "title": "Concrete Problems in AI Safety",
   "discipline": "philosophy",
   "year": 2016,
   "authors": "Dario Amodei, Chris Olah, Jacob Steinhardt, Paul Christiano, John Schulman, Dan Mané",
   "url": "papers/amodei2016-ai-safety.html",
   "key_terms": [
    "safety",
    "accidents",
    "objective",
    "learning",
    "reward",
    "exploration",
    "systems",
    "robustness",
    "supervision"
   ],
   "phenomenon": "AI safety (accident risks in ML: side effects, reward hacking, safe exploration)",
   "claim_type": "existence",
   "method": "theory",
   "statistics": [
    {
     "term": "Bayesian",
     "evidence": "“Bayesian optimization with safety con- straints: safe and automatic parameter tuning in robotics.” arXiv, 2016”.",
     "decision": "rejected",
     "role": "bibliography",
     "reason": "occurrence is confined to a citation/reference rather than the article body",
     "provenance": {
      "source_path": "papers/amodei2016-ai-safety.html",
      "source_body_sha256": "6984b0761fbd71111dae4ccc014cf9a9e225474ef448d31042dba83e5f23ae7b",
      "evidence_sha256": "abfe0f762052f108d2145fff97a42efba1fd1433dcb537b8d24223f7bd308b5c",
      "grounding_status": "verified-reference"
     }
    },
    {
     "term": "bootstrap",
     "evidence": "In the latter case, there has been interesting recent work on using bootstrapping to estimate finite-sample variation in the learned parameters of a neural network [114]; it seems ",
     "decision": "accepted",
     "role": "substantive-mention",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/amodei2016-ai-safety.html",
      "source_body_sha256": "6984b0761fbd71111dae4ccc014cf9a9e225474ef448d31042dba83e5f23ae7b",
      "evidence_sha256": "db3722a1b8d186091a49443c87eb435dfbc91c4f8d2c2378af6d43f6c8095414",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "correlation",
     "evidence": "igner chooses an objective function that is seemingly highly correlated with accomplishing the task, but that correlation breaks down when the objective function is being strongly ",
     "decision": "rejected",
     "role": "non-statistical-sense",
     "reason": "lexical match does not denote the controlled statistical concept",
     "provenance": {
      "source_path": "papers/amodei2016-ai-safety.html",
      "source_body_sha256": "6984b0761fbd71111dae4ccc014cf9a9e225474ef448d31042dba83e5f23ae7b",
      "evidence_sha256": "11f5b019e51c37faa4872d2dc5175784a6eae1bca503a0d21d42e07d3a10157b",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "hypothesis-testing",
     "evidence": "that are potentially relevant to this problem, including change detection and anomaly detection [21, 80, 91], hypothesis testing [145], transfer learning [138, 124, 125, 25], and s",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit hypothesis-testing language",
     "provenance": {
      "source_path": "papers/amodei2016-ai-safety.html",
      "source_body_sha256": "6984b0761fbd71111dae4ccc014cf9a9e225474ef448d31042dba83e5f23ae7b",
      "evidence_sha256": "1947baa40b5b01ced23c63d83fe5f6dfb029238f799e159a46adea088da98d1d",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "regression",
     "evidence": "For a simple example, consider a variant of linear regression where we might assume that y = (cid:104)w∗, x(cid:105) + v, where E[v|x] = 0, but we don’t make any further assumption",
     "decision": "accepted",
     "role": "substantive-mention",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/amodei2016-ai-safety.html",
      "source_body_sha256": "6984b0761fbd71111dae4ccc014cf9a9e225474ef448d31042dba83e5f23ae7b",
      "evidence_sha256": "c2cd85200a86c9947b7e2b689c406bd7a185a8529b2284cf37dab8d8519d7053",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "variance",
     "evidence": "This approach is limited by the variance of the importance estimate, which is very large or even infinite unless p0 and p∗ are close together.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/amodei2016-ai-safety.html",
      "source_body_sha256": "6984b0761fbd71111dae4ccc014cf9a9e225474ef448d31042dba83e5f23ae7b",
      "evidence_sha256": "41326fd6249b81220babb6cab6b489084f807d23a207d6c29550b691dd1ad108",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "arXiv-nonexclusive-distrib-1.0",
   "license_url": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
   "source_url": "https://arxiv.org/abs/1606.06565",
   "doi": "",
   "full_text_redistributable": false
  },
  {
   "id": "grilli2016-ecological-modularity",
   "title": "Modularity and Stability in Ecological Communities",
   "discipline": "ecology",
   "year": 2016,
   "authors": "Jacopo Grilli, Tim Rogers, Stefano Allesina",
   "url": "papers/grilli2016-ecological-modularity.html",
   "key_terms": [
    "modularity",
    "stability",
    "ecological",
    "community",
    "species",
    "network",
    "interactions",
    "eigenvalue",
    "matrix"
   ],
   "phenomenon": "stability of large ecological communities (effect of network modularity)",
   "claim_type": "mechanism",
   "method": "theory",
   "statistics": [
    {
     "term": "correlation",
     "evidence": "riate distribution (in this case, a bivariate normal with identical marginals defined by m 14 0, s 14 1/2 and correlation r 14 (cid:2) 3/4; red stands for negative coefficients, bl",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/grilli2016-ecological-modularity.html",
      "source_body_sha256": "b34b933954e211184f0b7c98d77767948ecd53cef57cb64a37b2268251559bee",
      "evidence_sha256": "3a565e748414bf51cdce9cd6792944962eb1c5f5818a2582de6de752c66b5048",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "mean",
     "evidence": "This directionality proved important in our previous study36, where we showed that when the mean of the negative effects dominates that of the positive effects, systems built accor",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/grilli2016-ecological-modularity.html",
      "source_body_sha256": "b34b933954e211184f0b7c98d77767948ecd53cef57cb64a37b2268251559bee",
      "evidence_sha256": "38022dbc91df53aa2b8f9d5d68078cc82eeb0e53ef89332fd5a4c99d581e4975",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "variance",
     "evidence": "14 0.1 in this case), correlation (r 14 (cid:2) 3/4) and size of the smaller subsystem (a 14 1/4), we fix the variance s2 14 1 and study three different cases, in which the mean m ",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/grilli2016-ecological-modularity.html",
      "source_body_sha256": "b34b933954e211184f0b7c98d77767948ecd53cef57cb64a37b2268251559bee",
      "evidence_sha256": "c634e3d24e8f440a308fd952d83d814d04b39dd0540bd73fb9c6e7854ea0c44c",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "CC-BY-4.0",
   "license_url": "http://creativecommons.org/licenses/by/4.0/",
   "source_url": "https://www.nature.com/articles/ncomms12031",
   "doi": "10.1038/ncomms12031",
   "full_text_redistributable": true
  },
  {
   "id": "henrich2010-weird",
   "title": "The Weirdest People in the World?",
   "discipline": "anthropology",
   "year": 2010,
   "authors": "Joseph Henrich, Steven J. Heine, Ara Norenzayan",
   "url": "papers/henrich2010-weird.html",
   "key_terms": [
    "WEIRD",
    "behavioral sciences",
    "cross-cultural",
    "generalizability",
    "population variability",
    "human universals",
    "external validity",
    "cultural psychology",
    "sampling",
    "representative"
   ],
   "phenomenon": "external validity / generalizability of behavioral findings (WEIRD sampling bias)",
   "claim_type": "critique",
   "method": "comparative",
   "statistics": [
    {
     "term": "correlation",
     "evidence": "Pearson.",
     "decision": "accepted",
     "role": "substantive-mention",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/henrich2010-weird.html",
      "source_body_sha256": "2462121759b02a995c365f3d4cedd3a81f046656d7ecff4d7fb9347701ad9385",
      "evidence_sha256": "3baa03bd987513249698aa8651a196d18940b92e14eec72d5dee88a3a5181485",
      "grounding_status": "verified-reference"
     }
    },
    {
     "term": "effect-size",
     "evidence": "For example, the cognitive ability to estimate quan- tity approximately, discussed earlier, appears to be quite invariant in that it produces cognitive responses with iden- tical e",
     "decision": "accepted",
     "role": "substantive-mention",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/henrich2010-weird.html",
      "source_body_sha256": "2462121759b02a995c365f3d4cedd3a81f046656d7ecff4d7fb9347701ad9385",
      "evidence_sha256": "d0edd90613449012a55c0e589f6194ae6f8e07bce6426b3ad50de930bbfb1b74",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "hypothesis-testing",
     "evidence": "Although rigorous science is always based on the refinement of models through hypothesis testing, one has to begin somewhere, and I strongly suspect that social scien- tists in gen",
     "decision": "rejected",
     "role": "non-statistical-sense",
     "reason": "lexical match does not denote the controlled statistical concept",
     "provenance": {
      "source_path": "papers/henrich2010-weird.html",
      "source_body_sha256": "2462121759b02a995c365f3d4cedd3a81f046656d7ecff4d7fb9347701ad9385",
      "evidence_sha256": "650cf649eba4fa9764d00f920fc1b06341a067b06b18fde993655ef270f062e6",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "mean",
     "evidence": "Research by Flynn (1987; 2007) shows that IQ scores increased over the last half century by an average of 18 points across all industrialized nations for which there were adequate ",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/henrich2010-weird.html",
      "source_body_sha256": "2462121759b02a995c365f3d4cedd3a81f046656d7ecff4d7fb9347701ad9385",
      "evidence_sha256": "0cd23ce61e7b06a58b6e10a3dede07f0ca44a8d6ed507691a283ddf26217b173",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "sample-size",
     "evidence": "he Internet samples may seem modest, Internet methods permit the collection of large samples, so the absolute sample sizes of non-Weird partici- pants can be quite impressive.",
     "decision": "accepted",
     "role": "substantive-mention",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/henrich2010-weird.html",
      "source_body_sha256": "2462121759b02a995c365f3d4cedd3a81f046656d7ecff4d7fb9347701ad9385",
      "evidence_sha256": "29ff1631df516d3f40e6d82409125022c93ee20b7dcc9da4462ad9e1fc8aad06",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "statistical-significance",
     "evidence": "While the San’s PSE value cannot be distinguished from zero, the American under- graduates’ PSE value is significantly different from all the other societies studied.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/henrich2010-weird.html",
      "source_body_sha256": "2462121759b02a995c365f3d4cedd3a81f046656d7ecff4d7fb9347701ad9385",
      "evidence_sha256": "3c86aa55d373bb8707ca22f442264076f0255d4381240737319bd83630c3023f",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "variance",
     "evidence": "The socio-ecological approach turns variance among populations from a liability to an asset doi:10.1017/S0140525X10000129 Selin Kesebir, Shigehiro Oishi, and Barbara A.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/henrich2010-weird.html",
      "source_body_sha256": "2462121759b02a995c365f3d4cedd3a81f046656d7ecff4d7fb9347701ad9385",
      "evidence_sha256": "477353db9a58c60886a5562bf56d4627ab65ebc3c9bbdb2131ee0ddbf5be2f82",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "open-access",
   "license_url": "",
   "source_url": "https://www2.psych.ubc.ca/~henrich/pdfs/WeirdPeople.pdf",
   "doi": "10.1017/S0140525X0999152X",
   "full_text_redistributable": false
  },
  {
   "id": "moher2009-prisma",
   "title": "Preferred Reporting Items for Systematic Reviews and Meta-Analyses: The PRISMA Statement",
   "discipline": "medicine",
   "year": 2009,
   "authors": "David Moher, Alessandro Liberati, Jennifer Tetzlaff, Douglas G. Altman",
   "url": "papers/moher2009-prisma.html",
   "key_terms": [
    "PRISMA",
    "systematic",
    "review",
    "meta-analysis",
    "reporting",
    "checklist",
    "transparent",
    "studies",
    "evidence"
   ],
   "phenomenon": "transparent reporting of systematic reviews (reducing bias in evidence synthesis)",
   "claim_type": "method",
   "method": "clinical-trial",
   "statistics": [
    {
     "term": "confidence-interval",
     "evidence": "For all outcomes considered (benefits or harms), present, for each study: (a) simple summary data for each intervention group and (b) effect estimates and confidence intervals, ide",
     "decision": "accepted",
     "role": "substantive-mention",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/moher2009-prisma.html",
      "source_body_sha256": "a586bfcf24e8993fd57c6b8d1cd03f6d8168890d0f81bc8c00e4ad4ed686fc76",
      "evidence_sha256": "95f2f1bb0d717ce4addaed6cd5a56a9b5126de83a5d684c30309119fce3741df",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "mean",
     "evidence": "Reporting was generally poor; between one and 14 characteristics were adequately reported (mean = 7.7; standard deviation = 2.7).",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "numeric mean or average in main text",
     "provenance": {
      "source_path": "papers/moher2009-prisma.html",
      "source_body_sha256": "a586bfcf24e8993fd57c6b8d1cd03f6d8168890d0f81bc8c00e4ad4ed686fc76",
      "evidence_sha256": "c36cc0d217393af638f95fc9a990eaccaef849342b40f41eccd1c32f82f8a771",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "standard-deviation",
     "evidence": "Reporting was generally poor; between one and 14 characteristics were adequately reported (mean = 7.7; standard deviation = 2.7).",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/moher2009-prisma.html",
      "source_body_sha256": "a586bfcf24e8993fd57c6b8d1cd03f6d8168890d0f81bc8c00e4ad4ed686fc76",
      "evidence_sha256": "c36cc0d217393af638f95fc9a990eaccaef849342b40f41eccd1c32f82f8a771",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "CC-BY-4.0",
   "license_url": "http://creativecommons.org/licenses/by/4.0/",
   "source_url": "https://journals.plos.org/plosmedicine/article?id=10.1371/journal.pmed.1000097",
   "doi": "10.1371/journal.pmed.1000097",
   "full_text_redistributable": true
  },
  {
   "id": "blei2003-lda",
   "title": "Latent Dirichlet Allocation",
   "discipline": "digital-humanities",
   "year": 2003,
   "authors": "David M. Blei, Andrew Y. Ng, Michael I. Jordan",
   "url": "papers/blei2003-lda.html",
   "key_terms": [
    "topic",
    "document",
    "corpus",
    "mixture",
    "probabilistic",
    "words",
    "generative",
    "Dirichlet",
    "variational"
   ],
   "phenomenon": "topic modeling (latent Dirichlet allocation; discovering topics in text)",
   "claim_type": "method",
   "method": "computational-model",
   "statistics": [
    {
     "term": "Bayesian",
     "evidence": "LDA is a three-level hierarchical Bayesian model, in which each item of a collection is modeled as a finite mixture over an underlying set of topics.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/blei2003-lda.html",
      "source_body_sha256": "83885f8ba8a6a746faccc0089c5507c2a6da71d6ec43d12f39a7b31cb1dc38d3",
      "evidence_sha256": "a845654a4c83b067fade2be3c1ba6788edfdbf9b3c4d69ba661683c2e2b9374a",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "mean",
     "evidence": "Laplace smoothing is commonly used; this essentially yields the mean of the posterior distribution under a uniform Dirichlet prior on the multinomial parameters.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/blei2003-lda.html",
      "source_body_sha256": "83885f8ba8a6a746faccc0089c5507c2a6da71d6ec43d12f39a7b31cb1dc38d3",
      "evidence_sha256": "d2ef24fae356fdb5c84c264485dc24bc30c83547826ff9b132d0f6cb75b08992",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "statistical-significance",
     "evidence": "We aim to demonstrate in the current paper that, by taking the de Finetti theorem seriously, we can capture significant intra-document statistical structure via the mixing distribu",
     "decision": "rejected",
     "role": "lexical-significance",
     "reason": "significant/significantly is used without inferential-statistical meaning",
     "provenance": {
      "source_path": "papers/blei2003-lda.html",
      "source_body_sha256": "83885f8ba8a6a746faccc0089c5507c2a6da71d6ec43d12f39a7b31cb1dc38d3",
      "evidence_sha256": "aa78ae78003be8dfbcdc32e30b7de5db29146cb3e14c64e7e438881495c3f702",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "open-access",
   "license_url": "",
   "source_url": "https://jmlr.org/papers/v3/blei03a.html",
   "doi": "",
   "full_text_redistributable": false
  },
  {
   "id": "elgammal2017-creative-gan",
   "title": "CAN: Creative Adversarial Networks, Generating Art by Learning About Styles and Deviating from Style Norms",
   "discipline": "art-history",
   "year": 2017,
   "authors": "Ahmed Elgammal, Bingchen Liu, Mohamed Elhoseiny, Marian Mazzone",
   "url": "papers/elgammal2017-creative-gan.html",
   "key_terms": [
    "art",
    "style",
    "creative",
    "GAN",
    "arousal",
    "novelty",
    "generated",
    "discriminator",
    "deviating"
   ],
   "phenomenon": "computational art generation (creativity via deviating from style norms)",
   "claim_type": "method",
   "method": "computational-model",
   "statistics": [
    {
     "term": "hypothesis-testing",
     "evidence": "2.8) We performed two-sample t-test to determine the statistical significance of these results, with the null hypothesis being that the subjects’ responses for both CAN and GAN are",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit hypothesis-testing language",
     "provenance": {
      "source_path": "papers/elgammal2017-creative-gan.html",
      "source_body_sha256": "1797c074c1476904b0ff51d200252a25ea497e85e8866f66392cc0e72ac70386",
      "evidence_sha256": "464be0bb39d92aba70e6a56ffc9c0affc337165ab9f76a5ba43ed771e9eb07da",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "null-hypothesis",
     "evidence": "2.8) We performed two-sample t-test to determine the statistical significance of these results, with the null hypothesis being that the subjects’ responses for both CAN and GAN are",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit null-hypothesis language",
     "provenance": {
      "source_path": "papers/elgammal2017-creative-gan.html",
      "source_body_sha256": "1797c074c1476904b0ff51d200252a25ea497e85e8866f66392cc0e72ac70386",
      "evidence_sha256": "464be0bb39d92aba70e6a56ffc9c0affc337165ab9f76a5ba43ed771e9eb07da",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "p-value",
     "evidence": "DCGAN) p-value = 1.9932e − 15 ‡ Q2 t-test (CAN vs.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit p-value or inferential p notation",
     "provenance": {
      "source_path": "papers/elgammal2017-creative-gan.html",
      "source_body_sha256": "1797c074c1476904b0ff51d200252a25ea497e85e8866f66392cc0e72ac70386",
      "evidence_sha256": "2792cc3c99d86dce8cfd06ce90ca458a87004c300d726599bd5157b20e8d5387",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "sample-size",
     "evidence": "1: Input: mini-batch images x, matching label ˆc, number of training batch steps S 2: for n = 1 to S do 3: 4: 5: 6: 7: z ∼ N (0, 1)Z {Draw sample of random noise} ˆx ← G(z) {Forwar",
     "decision": "rejected",
     "role": "non-statistical-sense",
     "reason": "lexical match does not denote the controlled statistical concept",
     "provenance": {
      "source_path": "papers/elgammal2017-creative-gan.html",
      "source_body_sha256": "1797c074c1476904b0ff51d200252a25ea497e85e8866f66392cc0e72ac70386",
      "evidence_sha256": "c95041c0d6d7a356b558f1ecb74e2e10cd58e55b74ad423b57de441597f9aac5",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "standard-deviation",
     "evidence": "Initialization and Training: The weights were initialized from a zero-centered Normal distribu- tion with standard deviation 0.02.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/elgammal2017-creative-gan.html",
      "source_body_sha256": "1797c074c1476904b0ff51d200252a25ea497e85e8866f66392cc0e72ac70386",
      "evidence_sha256": "3c5dc85b254a835ce597f9097b4956f358070a63c71bfc1d8b1ad9112c4ac9a5",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "statistical-significance",
     "evidence": "2.8) We performed two-sample t-test to determine the statistical significance of these results, with the null hypothesis being that the subjects’ responses for both CAN and GAN are",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit inferential-statistical context",
     "provenance": {
      "source_path": "papers/elgammal2017-creative-gan.html",
      "source_body_sha256": "1797c074c1476904b0ff51d200252a25ea497e85e8866f66392cc0e72ac70386",
      "evidence_sha256": "464be0bb39d92aba70e6a56ffc9c0affc337165ab9f76a5ba43ed771e9eb07da",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "t-test",
     "evidence": "DCGAN) p-value = 1.9932e − 15 ‡ Q2 t-test (CAN vs.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/elgammal2017-creative-gan.html",
      "source_body_sha256": "1797c074c1476904b0ff51d200252a25ea497e85e8866f66392cc0e72ac70386",
      "evidence_sha256": "2792cc3c99d86dce8cfd06ce90ca458a87004c300d726599bd5157b20e8d5387",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "arXiv-nonexclusive-distrib-1.0",
   "license_url": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
   "source_url": "https://arxiv.org/abs/1706.07068",
   "doi": "",
   "full_text_redistributable": false
  },
  {
   "id": "garimella2018-political-discourse",
   "title": "Political Discourse on Social Media: Echo Chambers, Gatekeepers, and the Price of Bipartisanship",
   "discipline": "political-science",
   "year": 2018,
   "authors": "Kiran Garimella, Gianmarco De Francisci Morales, Aristides Gionis, Michael Mathioudakis",
   "url": "papers/garimella2018-political-discourse.html",
   "key_terms": [
    "echo chamber",
    "users",
    "political",
    "network",
    "Twitter",
    "opinions",
    "partisan",
    "discourse",
    "engagement"
   ],
   "phenomenon": "political echo chambers on social media",
   "claim_type": "existence",
   "method": "computational-model",
   "statistics": [
    {
     "term": "Bayesian",
     "evidence": "Birds of the Same Feather Tweet Together: Bayesian Ideal Point Estimation Using Twitter Data.",
     "decision": "rejected",
     "role": "bibliography",
     "reason": "occurrence is confined to a citation/reference rather than the article body",
     "provenance": {
      "source_path": "papers/garimella2018-political-discourse.html",
      "source_body_sha256": "4fcf2cc14057d2c594a9ea17ad7aaaf734b9eb66148963317a05bace677b9fb3",
      "evidence_sha256": "b765f5e32ef485c44f66958947db4ade9c30f835e8365121bd508a348d06adbf",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "correlation",
     "evidence": "he difference between the two groups of datasets is stark: produc- tion and consumption polarities are highly correlated for Political datasets, which means that users indeed tend ",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/garimella2018-political-discourse.html",
      "source_body_sha256": "4fcf2cc14057d2c594a9ea17ad7aaaf734b9eb66148963317a05bace677b9fb3",
      "evidence_sha256": "2621c8e188eab2bab74a99eb5ba75613a155b178b2304cd5e9314ddbcb8797ae",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "effect-size",
     "evidence": "Finally, their tweets are more appreciated, i.e., a higher fraction of their tweets receives a retweet, albeit the effect size is smaller in this case (figure not shown).",
     "decision": "accepted",
     "role": "substantive-mention",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/garimella2018-political-discourse.html",
      "source_body_sha256": "4fcf2cc14057d2c594a9ea17ad7aaaf734b9eb66148963317a05bace677b9fb3",
      "evidence_sha256": "3a0cdd236e82eec29807486f04b1736526973bbfb4dc61ac173f5a20082cfb64",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "p-value",
     "evidence": "A ✓ indicates that the corresponding feature is significantly higher for the group of the column (p < 0.001) for at least 4 of the 6 thresh- olds δ used, for most datasets.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit p-value or inferential p notation",
     "provenance": {
      "source_path": "papers/garimella2018-political-discourse.html",
      "source_body_sha256": "4fcf2cc14057d2c594a9ea17ad7aaaf734b9eb66148963317a05bace677b9fb3",
      "evidence_sha256": "f5eab740f90ebd328f504d1f7dd6541a0966b3f0735b8f2c4e646fcea7588f7e",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "statistical-significance",
     "evidence": "Table 2 (second column) summarizes the results for partisan users and lists the features for which the difference is statistically significant on a majority of the datasets.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit inferential-statistical context",
     "provenance": {
      "source_path": "papers/garimella2018-political-discourse.html",
      "source_body_sha256": "4fcf2cc14057d2c594a9ea17ad7aaaf734b9eb66148963317a05bace677b9fb3",
      "evidence_sha256": "bebcd22455d79da05692f9c8fd4aa5e98291fbef843ffb36cd4e55b18d217a1d",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "arXiv-nonexclusive-distrib-1.0",
   "license_url": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
   "source_url": "https://arxiv.org/abs/1801.01665",
   "doi": "",
   "full_text_redistributable": false
  },
  {
   "id": "reagan2016-emotional-arcs",
   "title": "The Emotional Arcs of Stories Are Dominated by Six Basic Shapes",
   "discipline": "literature",
   "year": 2016,
   "authors": "Andrew J. Reagan, Lewis Mitchell, Dilan Kiley, Christopher M. Danforth, Peter Sheridan Dodds",
   "url": "papers/reagan2016-emotional-arcs.html",
   "key_terms": [
    "emotional arcs",
    "stories",
    "sentiment analysis",
    "narrative",
    "plot",
    "six basic shapes",
    "digital humanities",
    "matrix decomposition",
    "unsupervised learning",
    "Project Gutenberg"
   ],
   "phenomenon": "emotional arcs of stories (six basic shapes; sentiment trajectories)",
   "claim_type": "existence",
   "method": "corpus-analysis",
   "statistics": [
    {
     "term": "mean",
     "evidence": "ster, a selection of the 20 most central books to a fully-connected network of books are shown along with the average of the emotional arc for all books in the cluster, along with ",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/reagan2016-emotional-arcs.html",
      "source_body_sha256": "53c59ab9f2eff4f8026a6a9a75daff89e19b42c4c78aa4264265ec9ef67e11b6",
      "evidence_sha256": "68eff8ae71f7ee1608e6af8b584d766674aaa993f3f4c83180c8bd806ebdba79",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "sample-size",
     "evidence": "The cluster ID is given by numbering the clusters in order of linkage starting at 0, with each individual book representing a cluster of size 1 such that the final cluster (all boo",
     "decision": "rejected",
     "role": "non-statistical-sense",
     "reason": "lexical match does not denote the controlled statistical concept",
     "provenance": {
      "source_path": "papers/reagan2016-emotional-arcs.html",
      "source_body_sha256": "53c59ab9f2eff4f8026a6a9a75daff89e19b42c4c78aa4264265ec9ef67e11b6",
      "evidence_sha256": "3294fbbee84eeb60fce595e40d44504a46d61e32fba700de5204633ae675c7ff",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "variance",
     "evidence": "05) (3) We use Ward’s method to generate a hierarchical clus- tering of stories, which proceeds by minimizing variance for a node k in the set of nodes , with distance func- tion D",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/reagan2016-emotional-arcs.html",
      "source_body_sha256": "53c59ab9f2eff4f8026a6a9a75daff89e19b42c4c78aa4264265ec9ef67e11b6",
      "evidence_sha256": "4077d1c979248d97211640c0cd3b52ada124fa6ac7ed37306b01ecc084bf5482",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "CC-BY-4.0",
   "license_url": "https://creativecommons.org/licenses/by/4.0",
   "source_url": "https://arxiv.org/abs/1606.07772",
   "doi": "10.1140/epjds/s13688-016-0093-1",
   "full_text_redistributable": true
  },
  {
   "id": "lashley1951-serial-order",
   "title": "The Problem of Serial Order in Behavior",
   "discipline": "neuroscience",
   "year": 1951,
   "authors": "Karl S. Lashley",
   "url": "papers/lashley1951-serial-order.html",
   "key_terms": [
    "serial order",
    "behavior",
    "syntax",
    "language",
    "movements",
    "sequence",
    "temporal",
    "associative",
    "hierarchies"
   ],
   "phenomenon": "serial order in behavior",
   "claim_type": "critique",
   "method": "theory",
   "statistics": [
    {
     "term": "mean",
     "evidence": "There is an average of 300 visual cells Bring into each central pathway.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "numeric mean or average in main text",
     "provenance": {
      "source_path": "papers/lashley1951-serial-order.html",
      "source_body_sha256": null,
      "evidence_sha256": "e15adbf18ddc0aa9254b5275a1dc6a52a2cdea9fd3ad3e985d4e0707f35f7877",
      "grounding_status": "source-unavailable"
     }
    }
   ],
   "license_spdx": "personal-copy",
   "source_url": "https://scholar.google.com/scholar?q=The%20Problem%20of%20Serial%20Order%20in%20Behavior",
   "doi": "",
   "full_text_redistributable": false
  },
  {
   "id": "stroop1935-interference",
   "title": "Studies of Interference in Serial Verbal Reactions",
   "discipline": "cognitive-psychology",
   "year": 1935,
   "authors": "J. Ridley Stroop",
   "url": "papers/stroop1935-interference.html",
   "key_terms": [
    "interference",
    "color naming",
    "word reading",
    "color stimuli",
    "word stimuli",
    "reaction times",
    "associative inhibition",
    "serial verbal reactions",
    "practice"
   ],
   "phenomenon": "interference between color naming and word reading (Stroop effect)",
   "claim_type": "existence",
   "method": "experiment",
   "statistics": [
    {
     "term": "mean",
     "evidence": "The correction changed the mean on the NCfFd test from 108.7t0 IX°-3 a>»d the mean of the NC test from 63.0 to 63.3.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "numeric mean or average in main text",
     "provenance": {
      "source_path": "papers/stroop1935-interference.html",
      "source_body_sha256": null,
      "evidence_sha256": "eb26764ab3ad39dbc0a92349f20d23c8c86ab018467b1f8c9a52bb0bb2f74efd",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "median",
     "evidence": "(The medians on the two tests are 61.9 and 110.4 seconds respectively.) The standard deviation is increased in approximately the same ratio from 10.8 to 18.8.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/stroop1935-interference.html",
      "source_body_sha256": null,
      "evidence_sha256": "bebe25e1327b933e5c9bf2290565a2142d934f1debe40e8a72054004050cf4f7",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "standard-deviation",
     "evidence": "Results: Table I gives the means (m), standard deviations (a), differences (D), probable error of the difference (PEd), and the reliability of the difference (DJPEd) for the whole ",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/stroop1935-interference.html",
      "source_body_sha256": null,
      "evidence_sha256": "586fcc5d64d17010ef836d30767ea89506a72a9fdd2d78fb6392411831b30a40",
      "grounding_status": "source-unavailable"
     }
    }
   ],
   "license_spdx": "personal-copy",
   "source_url": "https://doi.org/",
   "doi": "10.1037/h0054651",
   "full_text_redistributable": false
  },
  {
   "id": "scarborough1977-lexical",
   "title": "Frequency and Repetition Effects in Lexical Memory",
   "discipline": "cognitive-psychology",
   "year": 1977,
   "authors": "Don L. Scarborough, Charles Cortese, Hollis S. Scarborough",
   "url": "papers/scarborough1977-lexical.html",
   "key_terms": [
    "word frequency",
    "lexical decision",
    "response time",
    "repetition",
    "lexical memory",
    "high-frequency",
    "low-frequency",
    "recency",
    "priming",
    "reaction time"
   ],
   "phenomenon": "repetition priming in lexical decision",
   "claim_type": "existence",
   "method": "experiment",
   "statistics": [
    {
     "term": "ANOVA",
     "evidence": "Responses to first and second presenta- tions of words were further analyzed in a six-way analysis of variance, with factors of response bias, presentation (first or second), case ",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/scarborough1977-lexical.html",
      "source_body_sha256": null,
      "evidence_sha256": "1132d7e2ec3ab939c0a14640abf89737d43437adf729645886d21e3fa4ae72dd",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "degrees-of-freedom",
     "evidence": "The RMSD from addi- tivity for the mean latencies for these two variables was 1.6 msec with df = 1.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/scarborough1977-lexical.html",
      "source_body_sha256": null,
      "evidence_sha256": "77ca78f0ed3a3ed281566f6b738580583d4e090c2f90147693264469e45e3f71",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "hypothesis-testing",
     "evidence": "Experiment 2 was designed to test this hypothesis.",
     "decision": "rejected",
     "role": "non-statistical-sense",
     "reason": "lexical match does not denote the controlled statistical concept",
     "provenance": {
      "source_path": "papers/scarborough1977-lexical.html",
      "source_body_sha256": null,
      "evidence_sha256": "1b1422b125333245f988bf0b76a47b060335899e523090151d820f20648587f1",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "mean",
     "evidence": "For high- frequency words, each word averaged 6.2 neighbors as compared with an average of 7.9 neighbors for low-frequency words, which is a nonsignificant difference, £(66) = 0.4,",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "numeric mean or average in main text",
     "provenance": {
      "source_path": "papers/scarborough1977-lexical.html",
      "source_body_sha256": null,
      "evidence_sha256": "b9ff60062386d2b3515f4e517950ae628e08d17359f7b5469f4dc0df3aef6254",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "p-value",
     "evidence": "cant, F'mln(l, 57) = 101, p < .01.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit p-value or inferential p notation",
     "provenance": {
      "source_path": "papers/scarborough1977-lexical.html",
      "source_body_sha256": null,
      "evidence_sha256": "b1044a03267f48cc78176090450189fe64ddb926401d4e3612c41a317b490d41",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "statistical-significance",
     "evidence": "For the nonwords, as with the words, there was a significant effect of repetition, .F'min (1, 178) = 15.0, p < .01.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/scarborough1977-lexical.html",
      "source_body_sha256": null,
      "evidence_sha256": "04b928ff02f8b6d0a968d49cd049c805c118a82414f38a18dee340e435b66b29",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "variance",
     "evidence": "For words, there was no significant effect of the number of trials intervening between presentations, although a failure to find a lag effect may simply be due to the small size of",
     "decision": "rejected",
     "role": "non-statistical-sense",
     "reason": "lexical match does not denote the controlled statistical concept",
     "provenance": {
      "source_path": "papers/scarborough1977-lexical.html",
      "source_body_sha256": null,
      "evidence_sha256": "0cc6e0fe34774f2a73d858f6b5c533bfb2b9dd57aca57750ca9f4689d918faa2",
      "grounding_status": "source-unavailable"
     }
    }
   ],
   "license_spdx": "personal-copy",
   "source_url": "https://doi.org/",
   "doi": "10.1037/0096-1523.3.1.1",
   "full_text_redistributable": false
  },
  {
   "id": "bock1986-syntactic",
   "title": "Syntactic Persistence in Language Production",
   "discipline": "linguistics",
   "year": 1986,
   "authors": "J. Kathryn Bock",
   "url": "papers/bock1986-syntactic.html",
   "key_terms": [
    "syntactic priming",
    "language production",
    "activation",
    "syntactic form",
    "sentence production",
    "persistence",
    "priming",
    "speech errors",
    "productivity",
    "structural"
   ],
   "phenomenon": "syntactic priming (structural persistence in language production)",
   "claim_type": "existence",
   "method": "experiment",
   "statistics": [
    {
     "term": "confidence-interval",
     "evidence": "Confidence intervals (shown in Table 1) for the differences between the means were constructed using the appro- priate values of the t statistic.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/bock1986-syntactic.html",
      "source_body_sha256": null,
      "evidence_sha256": "d6edba27579483b23f8cb7999267a858873215d26a9e49479abc4ba85d085fb2",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "correlation",
     "evidence": "escriptions of pictures with human versus nonhuman agents showed that the use of passive sentences was highly correlated with non- human agency.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/bock1986-syntactic.html",
      "source_body_sha256": null,
      "evidence_sha256": "ffc498b9a1cabbd2f9fb433083df871ab111307bc4382c9ea126367901766eaa",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "hypothesis-testing",
     "evidence": "An alternative hypothesis is that it is not type- or token-based sentence representations that are responsible for the repetition effect, but the syn- tactic processes responsible ",
     "decision": "rejected",
     "role": "non-statistical-sense",
     "reason": "lexical match does not denote the controlled statistical concept",
     "provenance": {
      "source_path": "papers/bock1986-syntactic.html",
      "source_body_sha256": null,
      "evidence_sha256": "277cee8d2785f484671a60db1dafca7c4e9e5fe80e0d4a88794343162237e236",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "p-value",
     "evidence": "All of these differences were significant for both subjects and items (p < .05).",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit p-value or inferential p notation",
     "provenance": {
      "source_path": "papers/bock1986-syntactic.html",
      "source_body_sha256": null,
      "evidence_sha256": "453cc6a86f01cabcf23cff93de3ae8104fc6a3372b5f4bde87637aff42cd97fa",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "statistical-significance",
     "evidence": "All of these differences were significant for both subjects and items (p < .05).",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/bock1986-syntactic.html",
      "source_body_sha256": null,
      "evidence_sha256": "453cc6a86f01cabcf23cff93de3ae8104fc6a3372b5f4bde87637aff42cd97fa",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "t-test",
     "evidence": "Confidence intervals (shown in Table 1) for the differences between the means were constructed using the appro- priate values of the t statistic.",
     "decision": "rejected",
     "role": "non-statistical-sense",
     "reason": "lexical match does not denote the controlled statistical concept",
     "provenance": {
      "source_path": "papers/bock1986-syntactic.html",
      "source_body_sha256": null,
      "evidence_sha256": "d6edba27579483b23f8cb7999267a858873215d26a9e49479abc4ba85d085fb2",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "variance",
     "evidence": "None of the in- teractions involving the prime agency factor were significant, with all Fs < 1 in analyses of variance treating subjects as random effects, and all Fs < 1.07 in ana",
     "decision": "rejected",
     "role": "non-statistical-sense",
     "reason": "lexical match does not denote the controlled statistical concept",
     "provenance": {
      "source_path": "papers/bock1986-syntactic.html",
      "source_body_sha256": null,
      "evidence_sha256": "a20a17f3fdc6e9f4f591506ef1e0c5691cdfb310e97ab686674c2b958e2a24a2",
      "grounding_status": "source-unavailable"
     }
    }
   ],
   "license_spdx": "personal-copy",
   "source_url": "https://doi.org/",
   "doi": "10.1016/0010-0285(86)90004-6",
   "full_text_redistributable": false
  },
  {
   "id": "fazio1986-attitudes",
   "title": "On the Automatic Activation of Attitudes",
   "discipline": "social-psychology",
   "year": 1986,
   "authors": "Russell H. Fazio, David M. Sanbonmatsu, Martha C. Powell, Frank R. Kardes",
   "url": "papers/fazio1986-attitudes.html",
   "key_terms": [],
   "phenomenon": "",
   "claim_type": "",
   "method": "experiment",
   "statistics": [
    {
     "term": "ANOVA",
     "evidence": "A 2 (strength of association) X 2 (prime valence) X 2 (target valence) analysis of variance was performed on the facilitation scores.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/fazio1986-attitudes.html",
      "source_body_sha256": null,
      "evidence_sha256": "d19548f834af8820f654e8ac3ce50b0b00cf2c28ce123d75632735c0ae59dbae",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "p-value",
     "evidence": "The analysis revealed that the expected three-way inter- action was statistically significant, F(l, 21 ) = 6.86, p < .02.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit p-value or inferential p notation",
     "provenance": {
      "source_path": "papers/fazio1986-attitudes.html",
      "source_body_sha256": null,
      "evidence_sha256": "b8da80b3004d693c6406cee81140beb412c30bf52784c93dead632f2c2fd7663",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "statistical-significance",
     "evidence": "The analysis revealed that the expected three-way inter- action was statistically significant, F(l, 21 ) = 6.86, p < .02.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit inferential-statistical context",
     "provenance": {
      "source_path": "papers/fazio1986-attitudes.html",
      "source_body_sha256": null,
      "evidence_sha256": "b8da80b3004d693c6406cee81140beb412c30bf52784c93dead632f2c2fd7663",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "variance",
     "evidence": "A 2 (strength of association) X 2 (prime valence) X 2 (target valence) analysis of variance was performed on the facilitation scores.",
     "decision": "rejected",
     "role": "non-statistical-sense",
     "reason": "lexical match does not denote the controlled statistical concept",
     "provenance": {
      "source_path": "papers/fazio1986-attitudes.html",
      "source_body_sha256": null,
      "evidence_sha256": "d19548f834af8820f654e8ac3ce50b0b00cf2c28ce123d75632735c0ae59dbae",
      "grounding_status": "source-unavailable"
     }
    }
   ],
   "license_spdx": "personal-copy",
   "source_url": "https://doi.org/",
   "doi": "10.1037/0022-3514.50.2.229",
   "full_text_redistributable": false
  },
  {
   "id": "logan1990-automaticity",
   "title": "Repetition Priming and Automaticity: Common Underlying Mechanisms?",
   "discipline": "cognitive-psychology",
   "year": 1990,
   "authors": "Gordon D. Logan",
   "url": "papers/logan1990-automaticity.html",
   "key_terms": [
    "repetition priming",
    "automaticity",
    "implicit memory",
    "power function",
    "instances",
    "lexical decision",
    "attention",
    "item specificity",
    "exposures",
    "learning"
   ],
   "phenomenon": "repetition priming and automaticity (instance-based retrieval)",
   "claim_type": "mechanism",
   "method": "theory",
   "statistics": [
    {
     "term": "ANOVA",
     "evidence": "These effects were confirmed by ANOVA: The main effect of word vs nonword was significant, F(1,50) = 119.43, p --c .Ol, MS, = 3220.48, as was the m",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/logan1990-automaticity.html",
      "source_body_sha256": null,
      "evidence_sha256": "c54a31f230572c0bea5fe8c7b85cf10e5d1ee2abf5b0fbc5002346f1470d5cbe",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "correlation",
     "evidence": "Error rate averaged 4% and was positively correlated with mean reaction time.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/logan1990-automaticity.html",
      "source_body_sha256": null,
      "evidence_sha256": "ad9db0b4b5427d8c5fed42bab7f91b5bfa2df3fed2b28034755b08b5f805bfe3",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "hypothesis-testing",
     "evidence": "three empirical parallels between repetition priming and automaticity predicted by the instance theory and to test hypotheses about possible associative bases of repeti- tion primi",
     "decision": "rejected",
     "role": "non-statistical-sense",
     "reason": "lexical match does not denote the controlled statistical concept",
     "provenance": {
      "source_path": "papers/logan1990-automaticity.html",
      "source_body_sha256": null,
      "evidence_sha256": "0ba9489e04d05498cd1457161b833d9e6bdc2aaf23bad019fc3ab8f05676044c",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "mean",
     "evidence": "The words were 340 common nouns selected from the Kucera and Francis (1967) frequency norms, ranging in frequency from 8 to 787 per million with a mean of 75.27.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "numeric mean or average in main text",
     "provenance": {
      "source_path": "papers/logan1990-automaticity.html",
      "source_body_sha256": null,
      "evidence_sha256": "0e4a68a14c60f6409d21c2991f49e6827340c812ccb6d519ccb76d941e160953",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "p-value",
     "evidence": "These effects were confirmed by ANOVA: The main effect of word vs nonword was significant, F(1,50) = 8.19, p < .Ol, MS, = 17881.50, as was the main effect of presentations, F(5,250",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/logan1990-automaticity.html",
      "source_body_sha256": null,
      "evidence_sha256": "7ecc7373d7cfc4de1e94ab096a8c17eccff73046d57c7a5d7db8ce91b39c9509",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "sample-size",
     "evidence": "Intuitively, the power- function speed-up follows from two opposing tendencies that govern the race: On the one hand, there are more opportunities to observe an ex- treme value as ",
     "decision": "rejected",
     "role": "non-statistical-sense",
     "reason": "lexical match does not denote the controlled statistical concept",
     "provenance": {
      "source_path": "papers/logan1990-automaticity.html",
      "source_body_sha256": null,
      "evidence_sha256": "469e1fcc4c3f776e7ad4e3eec4de37a6fb332f85ca9ebbd780f6570f37ae811e",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "standard-deviation",
     "evidence": "The instance theory predicts a power-function reduction in the standard deviation with practice, with the exponent of the function constrained to be equal to the exponent for the p",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/logan1990-automaticity.html",
      "source_body_sha256": null,
      "evidence_sha256": "83ad1e964d23c5e97861b1ea391047f1227d037e0be0f6d37ea3c82f6e8bb231",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "statistical-significance",
     "evidence": "These effects were confirmed by ANOVA: The main effect of word vs nonword was significant, F(1,50) = 119.43, p --c .Ol, MS, = 3220.48, as was the main effect of presentations, F(9,",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/logan1990-automaticity.html",
      "source_body_sha256": null,
      "evidence_sha256": "b329a0b31a1430075d8130b7f910c120d5305ab670d8639ebec1803eeaab05fa",
      "grounding_status": "source-unavailable"
     }
    }
   ],
   "license_spdx": "personal-copy",
   "source_url": "https://doi.org/",
   "doi": "10.1016/0010-0285(90)90002-L",
   "full_text_redistributable": false
  },
  {
   "id": "tulving-schacter1990-priming",
   "title": "Priming and Human Memory Systems",
   "discipline": "neuroscience",
   "year": 1990,
   "authors": "Endel Tulving, Daniel L. Schacter",
   "url": "papers/tulving-schacter1990-priming.html",
   "key_terms": [
    "priming",
    "memory systems",
    "implicit memory",
    "perceptual representation system",
    "semantic memory",
    "episodic memory",
    "procedural memory",
    "nonconscious",
    "identification",
    "dissociations"
   ],
   "phenomenon": "priming (perceptual representation system)",
   "claim_type": "mechanism",
   "method": "theory",
   "statistics": [
    {
     "term": "correlation",
     "evidence": "The results showed that with identical cues on both tests, the correlation between the tests was reliable (Q values around 0.90).",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/tulving-schacter1990-priming.html",
      "source_body_sha256": null,
      "evidence_sha256": "b79ed1e137b21e1ed6ccbdb3f2101228f4f45b1293aa7e85937353f2379c183b",
      "grounding_status": "source-unavailable"
     }
    }
   ],
   "license_spdx": "personal-copy",
   "source_url": "https://doi.org/",
   "doi": "10.1126/science.2296719",
   "full_text_redistributable": false
  },
  {
   "id": "bargh1996-social",
   "title": "Automaticity of Social Behavior: Direct Effects of Trait Construct and Stereotype Activation on Action",
   "discipline": "social-psychology",
   "year": 1996,
   "authors": "John A. Bargh, Mark Chen, Lara Burrows",
   "url": "papers/bargh1996-social.html",
   "key_terms": [
    "priming",
    "automaticity",
    "stereotype",
    "trait construct",
    "social behavior",
    "activation",
    "unconscious",
    "impression formation",
    "self-fulfilling prophecies",
    "social perception"
   ],
   "phenomenon": "automatic behavior priming (trait/stereotype activation influencing action)",
   "claim_type": "existence",
   "method": "experiment",
   "statistics": [
    {
     "term": "ANOVA",
     "evidence": "A one-way analysis of variance (ANOVA) of these data, with priming condition as the single factor, revealed a significant main effect, F(2, 33) = 5.76, p = .008.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/bargh1996-social.html",
      "source_body_sha256": null,
      "evidence_sha256": "703aad5b12a9e6b0b4e000e4e6d4154a1623df3733bcffaa140c7c460419cbc8",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "hypothesis-testing",
     "evidence": "The specific hypothesis tested was that the pro- portion of participants who interrupted the experimenter would increase as a function of the trait concept primed, from politeness ",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit hypothesis-testing language",
     "provenance": {
      "source_path": "papers/bargh1996-social.html",
      "source_body_sha256": null,
      "evidence_sha256": "c9fed514ae8ea100fd268c9f61a3d2d2e1801fbc443a02379553923f198bfc75",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "mean",
     "evidence": "We conducted a multivariate analysis of variance (MANOVA) using priming condition as the independent vari- able and the average of the four emotion scales and the average of the fo",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "numeric mean or average in main text",
     "provenance": {
      "source_path": "papers/bargh1996-social.html",
      "source_body_sha256": null,
      "evidence_sha256": "da456828044b829909ec0272b1752163aea6525ac4de00326056cf152088881f",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "p-value",
     "evidence": "A one-way analysis of variance (ANOVA) of these data, with priming condition as the single factor, revealed a significant main effect, F(2, 33) = 5.76, p = .008.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit p-value or inferential p notation",
     "provenance": {
      "source_path": "papers/bargh1996-social.html",
      "source_body_sha256": null,
      "evidence_sha256": "703aad5b12a9e6b0b4e000e4e6d4154a1623df3733bcffaa140c7c460419cbc8",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "statistical-significance",
     "evidence": "A one-way analysis of variance (ANOVA) of these data, with priming condition as the single factor, revealed a significant main effect, F(2, 33) = 5.76, p = .008.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/bargh1996-social.html",
      "source_body_sha256": null,
      "evidence_sha256": "703aad5b12a9e6b0b4e000e4e6d4154a1623df3733bcffaa140c7c460419cbc8",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "t-test",
     "evidence": "Participants in the elderly priming condition (M = 8.28 s) had a slower walking speed compared to participants in the neutral priming condi- tion (M = 7.30 s), t(28 ) = 2.86, p < .",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/bargh1996-social.html",
      "source_body_sha256": null,
      "evidence_sha256": "d4d266d5efee4baf19080177dd4f8c851b8a326b8102d36d023c3159afafba21",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "variance",
     "evidence": "A one-way analysis of variance (ANOVA) of these data, with priming condition as the single factor, revealed a significant main effect, F(2, 33) = 5.76, p = .008.",
     "decision": "rejected",
     "role": "non-statistical-sense",
     "reason": "lexical match does not denote the controlled statistical concept",
     "provenance": {
      "source_path": "papers/bargh1996-social.html",
      "source_body_sha256": null,
      "evidence_sha256": "703aad5b12a9e6b0b4e000e4e6d4154a1623df3733bcffaa140c7c460419cbc8",
      "grounding_status": "source-unavailable"
     }
    }
   ],
   "license_spdx": "personal-copy",
   "source_url": "https://doi.org/",
   "doi": "10.1037/0022-3514.71.2.230",
   "full_text_redistributable": false
  },
  {
   "id": "breitmeyer2005-priming",
   "title": "Unconscious and Conscious Priming",
   "discipline": "neuroscience",
   "year": 2005,
   "authors": "Bruno G. Breitmeyer",
   "url": "papers/breitmeyer2005-priming.html",
   "key_terms": [],
   "phenomenon": "",
   "claim_type": "",
   "method": "experiment",
   "statistics": [],
   "license_spdx": "personal-copy",
   "source_url": "https://scholar.google.com/scholar?q=Unconscious%20and%20Conscious%20Priming",
   "doi": "",
   "full_text_redistributable": false
  },
  {
   "id": "vohs2006-money",
   "title": "The Psychological Consequences of Money",
   "discipline": "economics",
   "year": 2006,
   "authors": "Kathleen D. Vohs, Nicole L. Mead, Miranda R. Goode",
   "url": "papers/vohs2006-money.html",
   "key_terms": [
    "money",
    "primed",
    "participants",
    "help",
    "reminders",
    "behavior",
    "experiments",
    "concept",
    "social"
   ],
   "phenomenon": "money priming (money cues shift behavior toward self-sufficiency)",
   "claim_type": "existence",
   "method": "experiment",
   "statistics": [
    {
     "term": "ANOVA",
     "evidence": "As predicted, participants who were reminded of money (play money and money prime) worked longer than control participants before requesting help [F(2,49) = 3.73,P <0 .",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/vohs2006-money.html",
      "source_body_sha256": null,
      "evidence_sha256": "6fe70678e87da1aaafd78ff75d500ba53dc5059e5e43d148090aad8aa82363f0",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "chi-square",
     "evidence": "In the last sentence of the fourth-to-last paragraph, “t(34)” should h ave been “χ2 (1)” .",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/vohs2006-money.html",
      "source_body_sha256": null,
      "evidence_sha256": "20be132ada40cc0e849ae1cd832a0567f32b0a52de29d268bdb84ac992b3db61",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "hypothesis-testing",
     "evidence": "To convincingly demonstrate that money makes people self-sufficient, we tested the hypothesis in new contexts.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit hypothesis-testing language",
     "provenance": {
      "source_path": "papers/vohs2006-money.html",
      "source_body_sha256": null,
      "evidence_sha256": "7defabb72743141438d9fbbc141444249e83b4f13d5989e84015a27298942df0",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "mean",
     "evidence": "Translated into time, control condition participants volunteered an average of 42.5 min of their time, whereas participants in the money condition volunteered only slightly more th",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "numeric mean or average in main text",
     "provenance": {
      "source_path": "papers/vohs2006-money.html",
      "source_body_sha256": null,
      "evidence_sha256": "48b90b6e0fe5cada932793e2d8863103ac91042420a4764c303a6a61ae1834bf",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "odds-ratio",
     "evidence": "duced among money condition participants rela- tive to participants in both the fish [X2(1) = 7.00, P < 0.05; odds ratio = 11.25] and no-screensaver conditions [X 2(1) = 8.22, P < ",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/vohs2006-money.html",
      "source_body_sha256": null,
      "evidence_sha256": "86e9405f092a76db1ef16ba82f3509d9ade94232b03193a12a4ba608e5536051",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "p-value",
     "evidence": "< 1], but each was significantly different from the control group [money prime versus control: t(49) = 2.44, P < 0.02; Cohen ’s d = 0.86; play money versus control: t(49) = 2.30, P",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit p-value or inferential p notation",
     "provenance": {
      "source_path": "papers/vohs2006-money.html",
      "source_body_sha256": null,
      "evidence_sha256": "8a2de4c7811ffa0e51e85cb23b7b121800d417050d4eb31135f3266605f1d442",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "sample-size",
     "evidence": "Participants’ desire to work with a peer was significantly affected by priming condition [X 2(2, n = 37) = 10.10, P < 0.01] (Table 2).",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/vohs2006-money.html",
      "source_body_sha256": null,
      "evidence_sha256": "ee47ce0b6dca7e5d807c1ae9fbd04992d7bab7e74b9933c5abbae886bdb2cc56",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "standard-deviation",
     "evidence": "0 4 ;m e a n(M)m o n e y prime = 314.06 s, SD = 172.79;M play money = 305.22 s, SD = 162.47;M control = 186.12 s, SD = 118.09].",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/vohs2006-money.html",
      "source_body_sha256": null,
      "evidence_sha256": "15dbdf53ce01fc8303478e5a45ca3d18d736a552eaece0401ac83f7a56cf7125",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "statistical-significance",
     "evidence": "The two money conditions did not differ from each other [ t(49) < 1], but each was significantly different from the control group [money prime versus control: t(49) = 2.44, P < 0.0",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/vohs2006-money.html",
      "source_body_sha256": null,
      "evidence_sha256": "655340ec3458f243120a28ef9488d3ecdd1e589f2539500aacd7be522e67ba70",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "t-test",
     "evidence": "other [ t(49) < 1], but each was significantly different from the control group [money prime versus control: t(49) = 2.44, P < 0.02; Cohen ’s d = 0.86; play money versus control: t",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/vohs2006-money.html",
      "source_body_sha256": null,
      "evidence_sha256": "1f3be01bb2bc2412872d397fa11f9389e597ee93b08eee278431ed9d03c6a2fd",
      "grounding_status": "source-unavailable"
     }
    }
   ],
   "license_spdx": "personal-copy",
   "source_url": "https://doi.org/",
   "doi": "10.1126/science.1132491",
   "full_text_redistributable": false
  },
  {
   "id": "mcmahon-olson2007-suppression",
   "title": "Repetition Suppression in Monkey Inferotemporal Cortex: Relation to Behavioral Priming",
   "discipline": "neuroscience",
   "year": 2007,
   "authors": "David B. T. McMahon, Carl R. Olson",
   "url": "papers/mcmahon-olson2007-suppression.html",
   "key_terms": [],
   "phenomenon": "",
   "claim_type": "",
   "method": "experiment",
   "statistics": [
    {
     "term": "ANOVA",
     "evidence": "To assess the contribution of specic experimental parame- ters to the priming effect, we conducted a three-way ANOVA on the magnitude of priming (RT 1 /H11002RT2) with lag, stimulu",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/mcmahon-olson2007-suppression.html",
      "source_body_sha256": null,
      "evidence_sha256": "12519c3e2306cd17a5740376452326ece63f2ad1033826ffaf61fc1c4199f3f5",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "chi-square",
     "evidence": "Both measures were signicantly biased in each monkey, and the bias was signicantly greater in m2 than in m1 ( P /H110210.05, chi-squared tests).",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/mcmahon-olson2007-suppression.html",
      "source_body_sha256": null,
      "evidence_sha256": "1e8f14f5006ded7e39a8a281c2bec52feeaebd859462223396c302573026239c",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "correlation",
     "evidence": "and suppression To assess whether neuronal suppression and behavioral priming were correlated, we performed a Pearson correlation analysis across all pairs of trials between measur",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/mcmahon-olson2007-suppression.html",
      "source_body_sha256": null,
      "evidence_sha256": "d0c20e7e701a1aabdcd78810a894386305b1c9cf2c2ec806864dee8dbb64a114",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "false-discovery-rate",
     "evidence": "Signicance of kernels was evaluated for each monkey and each lag by rank-sum tests, with Bonferroni correction.",
     "decision": "rejected",
     "role": "non-statistical-sense",
     "reason": "lexical match does not denote the controlled statistical concept",
     "provenance": {
      "source_path": "papers/mcmahon-olson2007-suppression.html",
      "source_body_sha256": null,
      "evidence_sha256": "bc7770064dab77732abd7fa3671ae1fb84f76951c17989bee3bbb1cb2df490a3",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "hypothesis-testing",
     "evidence": "Conversely, we can reject the hypothesis that there was a correlation in the experimental data set as strong as the simu- lated correlation.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/mcmahon-olson2007-suppression.html",
      "source_body_sha256": null,
      "evidence_sha256": "0c7215ccd9cd3d85647852dcaf20dc56a2383e7ed94675c337e0ee1de3da5ce5",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "mean",
     "evidence": "The function Pois generates a Poisson random number with a mean of Xn.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "numeric mean or average in main text",
     "provenance": {
      "source_path": "papers/mcmahon-olson2007-suppression.html",
      "source_body_sha256": null,
      "evidence_sha256": "8e03f6abd0908b3bcdb2b45da0156cc8e8ae6c711aa17204694cc0435a414e72",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "p-value",
     "evidence": "F-Statistic P-Value lag 5 5.61 /H11003104 1.12 /H11003104 2.5 0.0028 type 1 5.65 /H11003104 5.65 /H11003104 12.7 0.0004 monkey 1 1.75 /H11003104 1.75 /H11003104 3.9 0.048 lag /H11",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit p-value or inferential p notation",
     "provenance": {
      "source_path": "papers/mcmahon-olson2007-suppression.html",
      "source_body_sha256": null,
      "evidence_sha256": "28ed7a06871d3be22ae8cb7c3549e378e5535460661836fcfc5b551bd8d4d01d",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "regression",
     "evidence": "To answer these questions, we conducted a linear regression analysis on the visual responses evoked by rst and second presentations across all cases in which the monkey",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/mcmahon-olson2007-suppression.html",
      "source_body_sha256": null,
      "evidence_sha256": "234cf6dfd8fb20cb8bd64763e91669402c012ed882951aeb07314e05267cccb4",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "t-test",
     "evidence": "At each step, we used a two-sided, paired t-test to compare the ring rates evoked on the rst and second presentations of each stimulus.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/mcmahon-olson2007-suppression.html",
      "source_body_sha256": null,
      "evidence_sha256": "f14d361ed166e09400ef0c2c9d3f0cc26fbf22f22ce527a0c84ef5a65d484d99",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "variance",
     "evidence": "cant a correlation such that ring rate explained only a tiny fraction (0.026 2 /H110050.00068) of cross-trial variance in reaction time.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/mcmahon-olson2007-suppression.html",
      "source_body_sha256": null,
      "evidence_sha256": "3476afbdc945ff8582a76ea00e31eb265b7c30a4fdb0c92167df6afc868ecbc8",
      "grounding_status": "source-unavailable"
     }
    }
   ],
   "license_spdx": "personal-copy",
   "source_url": "https://doi.org/",
   "doi": "10.1152/jn.00160.2007",
   "full_text_redistributable": false
  },
  {
   "id": "aquino2009-moral",
   "title": "Testing a Social-Cognitive Model of Moral Behavior",
   "discipline": "social-psychology",
   "year": 2009,
   "authors": "Karl Aquino, Dan Freeman, Americus Reed II, Vivien K. G. Lim, Will Felps",
   "url": "papers/aquino2009-moral.html",
   "key_terms": [
    "moral identity",
    "moral behavior",
    "social-cognitive",
    "situational factors",
    "self-concept",
    "accessibility",
    "moral intentions",
    "prosocial behavior",
    "morality",
    "priming"
   ],
   "phenomenon": "moral identity accessibility (situational priming of the moral self-schema)",
   "claim_type": "mechanism",
   "method": "experiment",
   "statistics": [
    {
     "term": "ANOVA",
     "evidence": "We then conducted a repeated measures ANOVA on cooperation across the five trial blocks within each of these four groups.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/aquino2009-moral.html",
      "source_body_sha256": null,
      "evidence_sha256": "5f19b47cb15e754b8ed11b4c55727b41cc48e0dbc4af663ad81f3b37cd5aed1f",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "correlation",
     "evidence": "To create a general measure of the acces- sibility of achievement-oriented identity schemas, participants’ rankings of these two options were averaged (Spearman /H9267/H11005.36, p",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/aquino2009-moral.html",
      "source_body_sha256": null,
      "evidence_sha256": "cd55343ca04f534be7a88d6a20749aec5c900eca6fac732b90760911c7346e5c",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "hypothesis-testing",
     "evidence": "Hypothesis Tests One hundred dyads reached agreement and 12 impassed.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit hypothesis-testing language",
     "provenance": {
      "source_path": "papers/aquino2009-moral.html",
      "source_body_sha256": null,
      "evidence_sha256": "1ffff7436d2a8ab6097609a4c2bc39fd9ba84a8264399c278ea3f0860765105b",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "mean",
     "evidence": "There are 24 choices like this, with the property that across all alternative items, mean (self) /H11005mean (other) /H110050.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "numeric mean or average in main text",
     "provenance": {
      "source_path": "papers/aquino2009-moral.html",
      "source_body_sha256": null,
      "evidence_sha256": "0ef36701916cb492a5b4f447ca09be833debef8e5497b175582d9c22b5f53478",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "regression",
     "evidence": "2 The dependent variable consisted of four nonordered categorical responses, so we used multinomial logistic regression with maximum-likelihood estimation to conduct an initial tes",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/aquino2009-moral.html",
      "source_body_sha256": null,
      "evidence_sha256": "02ff7fd7d05daabcbbc973fb176d7e6c9f00a889e5b80f236d7d121a77b5f24c",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "standard-deviation",
     "evidence": "ition on the mediator (current acces- sibility of moral identity within the working self-concept) at /H110061 standard deviations of the moderator (centrality of moral identity) to",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/aquino2009-moral.html",
      "source_body_sha256": null,
      "evidence_sha256": "4edf989abd8afb444ba9a5274b6aa28553ded0ed85ffc1d6d15a5bb965b71cc3",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "statistical-significance",
     "evidence": "We con- ducted a formal test of whether this observed difference was statistically significant by creating two response categories: lie vs.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit inferential-statistical context",
     "provenance": {
      "source_path": "papers/aquino2009-moral.html",
      "source_body_sha256": null,
      "evidence_sha256": "3e107de7869849e6e97b07efbec6b487738d0fdb5118dff73d1ec8a3a4708f17",
      "grounding_status": "source-unavailable"
     }
    }
   ],
   "license_spdx": "personal-copy",
   "source_url": "https://doi.org/",
   "doi": "10.1037/a0015406",
   "full_text_redistributable": false
  },
  {
   "id": "ferraro2009-brand",
   "title": "The Power of Strangers: The Effect of Incidental Consumer Brand Encounters on Brand Choice",
   "discipline": "marketing",
   "year": 2009,
   "authors": "Rosellina Ferraro, James R. Bettman, Tanya L. Chartrand",
   "url": "papers/ferraro2009-brand.html",
   "key_terms": [],
   "phenomenon": "",
   "claim_type": "",
   "method": "experiment",
   "statistics": [
    {
     "term": "ANOVA",
     "evidence": "The four brands also did not differ(F(3, 44) ! 1.0, on general knowledge of the brand NS).(F(3, 44) p 1.73, There was a marginally signicant difference in",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/ferraro2009-brand.html",
      "source_body_sha256": null,
      "evidence_sha256": "c371b6d7684c7e3ffcc0ed5518c70f634c510aa1c8cdec387affe3ed57fe7e28",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "correlation",
     "evidence": "The strength of this correlation increased with ex- posure to Dasani relative to the control group in the no mis- attribution condition ( r p 0.47, r p 0.66, r p 0-exp 4-exp 12-exp",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/ferraro2009-brand.html",
      "source_body_sha256": null,
      "evidence_sha256": "7cafa1edaeea9495bf59e592dce2194e43f14bc29122b255cc01a595087c9291",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "mean",
     "evidence": "f group (i.e., 0 p out-group, 1 p in-group), and the value of construal at one standard de- viation below the mean (4 A), at the mean (4 B), and at one standard deviation above the",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "numeric mean or average in main text",
     "provenance": {
      "source_path": "papers/ferraro2009-brand.html",
      "source_body_sha256": null,
      "evidence_sha256": "d104a09c4c125945b172af9da8339757cc78ad164c67431d5d0293b3cb805cd2",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "regression",
     "evidence": "A logistic regression was used to analyze the data, with the dependent variable coded as 1 p chose Dasani or 0 p did not choose Dasani.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/ferraro2009-brand.html",
      "source_body_sha256": null,
      "evidence_sha256": "3e8817a0b928721c4101b7b3ff687ee00a6c751cc60a0ecc4fff142725f6a837",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "standard-deviation",
     "evidence": ", type of group (i.e., 0 p out-group, 1 p in-group), and the value of construal at one standard de- viation below the mean (4 A), at the mean (4 B), and at one standard deviation a",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/ferraro2009-brand.html",
      "source_body_sha256": null,
      "evidence_sha256": "ee827124d70f886d12f65e85d3221227ffc7a6175f2b1c5f4d7af00521d3cbed",
      "grounding_status": "source-unavailable"
     }
    }
   ],
   "license_spdx": "personal-copy",
   "source_url": "https://doi.org/",
   "doi": "10.1086/597331",
   "full_text_redistributable": false
  },
  {
   "id": "friis2017-nudge",
   "title": "Comparison of Three Nudge Interventions to Promote Vegetable Consumption",
   "discipline": "behavioral-economics",
   "year": 2017,
   "authors": "Rasmus Friis, Laurits R. Skov, et al.",
   "url": "papers/friis2017-nudge.html",
   "key_terms": [
    "nudge",
    "vegetable",
    "consumption",
    "priming",
    "default",
    "participants",
    "intervention",
    "choice architecture",
    "variety"
   ],
   "phenomenon": "nudging / choice architecture (nudge interventions on food choice)",
   "claim_type": "existence",
   "method": "experiment",
   "statistics": [
    {
     "term": "ANOVA",
     "evidence": "Oneway ANOVA with post hoc test was performed to explore any changes between the three nudges for: total intake, total chilli con carne intake, total rice intake and total vegetabl",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/friis2017-nudge.html",
      "source_body_sha256": null,
      "evidence_sha256": "c38eb94aa852b7ccd0cc93ffe5637c2e291c3b980cc5fe3a5d3c7af6f864753f",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "chi-square",
     "evidence": "2 With the use of the chi-square test.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/friis2017-nudge.html",
      "source_body_sha256": null,
      "evidence_sha256": "77676049fecc42428f66c4f2a4732e7443128f3cd22079cbc4f1468031fcd1a5",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "effect-size",
     "evidence": "Comparing the effect size of the three nudges (presented as the differ- ence between intervention and control in g) a significant diffe",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/friis2017-nudge.html",
      "source_body_sha256": null,
      "evidence_sha256": "1d7cd65c28fc87c474f5b5cfc8a9d70979e03c096d22d16847ec724b26de6c79",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "hypothesis-testing",
     "evidence": "The study had hoped to have a more representative sample, but it was not a requirement of the study’s design nor a theoretical basis for the alternative hypothesis’ expected effect",
     "decision": "rejected",
     "role": "non-statistical-sense",
     "reason": "lexical match does not denote the controlled statistical concept",
     "provenance": {
      "source_path": "papers/friis2017-nudge.html",
      "source_body_sha256": null,
      "evidence_sha256": "189513b6e250c3dd39bcde93b2a7176392adb4c13b65a1f304aab41eb375dc32",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "mean",
     "evidence": "3 Mean ± SD, all such values.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "numeric mean or average in main text",
     "provenance": {
      "source_path": "papers/friis2017-nudge.html",
      "source_body_sha256": null,
      "evidence_sha256": "506a345ad689cb14f3c4c03e6b4d2d3c701f8a069f308f408c0ec7378d721e2d",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "p-value",
     "evidence": "Results are presented as means ± SDs, with cor- responding p-values for comparisons for both the analysis within and between nudges.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit p-value or inferential p notation",
     "provenance": {
      "source_path": "papers/friis2017-nudge.html",
      "source_body_sha256": null,
      "evidence_sha256": "72da15ae1bda9d4e54b51fe231acf248d894617c98b87d15ec00e61681bd222d",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "standard-deviation",
     "evidence": "3 Mean ± SD, all such values.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/friis2017-nudge.html",
      "source_body_sha256": null,
      "evidence_sha256": "506a345ad689cb14f3c4c03e6b4d2d3c701f8a069f308f408c0ec7378d721e2d",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "statistical-significance",
     "evidence": "ction and appetite levels were mainly not significant within and between par- ticipants, and if they achieved statistical significance, the difference was not meaningful.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit inferential-statistical context",
     "provenance": {
      "source_path": "papers/friis2017-nudge.html",
      "source_body_sha256": null,
      "evidence_sha256": "fcc67f6ebd1d5a0b59cb63002ae0508da30da24fcf84e629df510227a29e9c05",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "t-test",
     "evidence": "Control Intervention 6.8±1.4 7.0±1.4 2.4±1.0 8.6±1.0 6.4±1.6 6.8±1.4 2.6±1.0 8.4±1.0 p* ns ns ns ns * Student t-test analysis of appetite scores from control day compared to interv",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/friis2017-nudge.html",
      "source_body_sha256": null,
      "evidence_sha256": "a836c3ba9cc5127b6038d757f46befe4e236c5dbb721447323ecc2cd7ececf4e",
      "grounding_status": "source-unavailable"
     }
    }
   ],
   "license_spdx": "personal-copy",
   "source_url": "https://doi.org/",
   "doi": "10.1371/journal.pone.0176028",
   "full_text_redistributable": false
  },
  {
   "id": "jitendra2016-math",
   "title": "Mathematical Representation and Schema-Based Instruction for Students with Learning Difficulties",
   "discipline": "education",
   "year": 2016,
   "authors": "Asha K. Jitendra, et al.",
   "url": "papers/jitendra2016-math.html",
   "key_terms": [
    "mathematics",
    "students",
    "representation",
    "schema",
    "instruction",
    "learning",
    "disabilities",
    "interventions",
    "achievement"
   ],
   "phenomenon": "schema-based instruction for mathematical word-problem solving",
   "claim_type": "method",
   "method": "experiment",
   "statistics": [
    {
     "term": "confidence-interval",
     "evidence": "h- quality or four acceptable-quality studies supporting the practice and the weighted effect size with a 95% confidence interval is significantly greater than zero; a practice is ",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/jitendra2016-math.html",
      "source_body_sha256": null,
      "evidence_sha256": "9adfd4846c011165907c9c80f8168ffa58b2a320251979c5900fe81840059e0f",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "correlation",
     "evidence": "ions (focus on relevant data and relations in the problem) are more mean- ingful, and their use is positively correlated with successful problem-solving perfor - mance (Hegarty & K",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/jitendra2016-math.html",
      "source_body_sha256": null,
      "evidence_sha256": "127510550758eac9e1a15a4c445d76c564e824228b9c37c0d3cfcfb31c19de80",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "effect-size",
     "evidence": "11 an effect size of 0.46 for teacher and student use of visual representations.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/jitendra2016-math.html",
      "source_body_sha256": null,
      "evidence_sha256": "76140e74bccb667ca5441ed217776db6f867cb0b2c7febeb9a6333a3526c4567",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "mean",
     "evidence": "We also used the WWC (2014) addi- tional guidelines to calculate effect sizes (e.g., when studies reported unadjusted group means, when studies used different pretest and posttest ",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/jitendra2016-math.html",
      "source_body_sha256": null,
      "evidence_sha256": "ea61fb4a545e95a2401662187ef9a9086574d740e4ea0845cb5fd726bbe4d071",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "sample-size",
     "evidence": "We then applied Hedges’ bias correction for small sample sizes to each of the effect sizes and calculated the weighted mean effect size and 95% confi- dence interval.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/jitendra2016-math.html",
      "source_body_sha256": null,
      "evidence_sha256": "15bd44737963c8cc8c72955e7961e9cd97e00bca094df42dbeb74cd16d06ac0f",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "standard-deviation",
     "evidence": "The study used an experimental or quasiexperimental design, and the study provided sufficient information (e.g., means, standard deviations, t or F statistics) to calculate effect ",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/jitendra2016-math.html",
      "source_body_sha256": null,
      "evidence_sha256": "553e789332583b7b227555a4d67c5235935656fcd6b3b027dd56275d86023ec1",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "standard-error",
     "evidence": "Weighted Effect Sizes, Standard Errors, and Confidence Intervals by Study Characteristics Across High- and Acceptable-Quality Studies.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit standard error or SEM",
     "provenance": {
      "source_path": "papers/jitendra2016-math.html",
      "source_body_sha256": null,
      "evidence_sha256": "057abeae50542aa5ca2aedca83bb127d2f9f60f8a77de4dd130da5202b5d4dde",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "statistical-significance",
     "evidence": "ptable-quality studies supporting the practice and the weighted effect size with a 95% confidence interval is significantly greater than zero; a practice is considered promising wh",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/jitendra2016-math.html",
      "source_body_sha256": null,
      "evidence_sha256": "d23a6a9e9550a12bb797582d71ca25a4399e0867a5b6903e9f91d33c8c38825c",
      "grounding_status": "source-unavailable"
     }
    }
   ],
   "license_spdx": "personal-copy",
   "source_url": "https://doi.org/",
   "doi": "10.1177/0014402915625062",
   "full_text_redistributable": false
  },
  {
   "id": "sivroni2023-auditory",
   "title": "Short-Term Auditory Priming in Freely-Moving Mice",
   "discipline": "neuroscience",
   "year": 2023,
   "authors": "Shir Sivroni, Hadas E. Sloin, Eran Stark",
   "url": "papers/sivroni2023-auditory.html",
   "key_terms": [],
   "phenomenon": "",
   "claim_type": "",
   "method": "experiment",
   "statistics": [
    {
     "term": "bootstrap",
     "evidence": "We quantied the ‘‘priming effect’’ in success rates by the difference between CPT and IPT success rates, which was 17% in the example session (p = 0.011, permutation test).",
     "decision": "rejected",
     "role": "non-statistical-sense",
     "reason": "lexical match does not denote the controlled statistical concept",
     "provenance": {
      "source_path": "papers/sivroni2023-auditory.html",
      "source_body_sha256": null,
      "evidence_sha256": "3943b0d9eb1c3bfbadb8afe12562e3acd70ea1388dea1ff1dc8702ef3968e1ae",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "effect-size",
     "evidence": "(E) The p values of the priming effect for every priming session (permutation test, comparing to an effect size of zero).",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/sivroni2023-auditory.html",
      "source_body_sha256": null,
      "evidence_sha256": "3a50543ae9929cf03ef60fb2c0b4f60837c9672a018144d0b495a97a722d9fb4",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "median",
     "evidence": "During the post-learning sessions, all subjects exhibited consistently high performance ( Figure 3 F), with a median [interquartile-interval, IQR] post- learning success rate of 88",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/sivroni2023-auditory.html",
      "source_body_sha256": null,
      "evidence_sha256": "0aff5ffdd6f9b40281b3522cfaf4c8df0ea7cdf907b3ba906700fc381cb9b774",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "p-value",
     "evidence": "(D) The p values of deviation from chance for subject mB278 sessions (Binomial test comparing to chance level, 0.5).",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit p-value or inferential p notation",
     "provenance": {
      "source_path": "papers/sivroni2023-auditory.html",
      "source_body_sha256": null,
      "evidence_sha256": "7b876446b0d287c85694fee51c3c876ace17460d8d5b5e91283103ada17370b5",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "sample-size",
     "evidence": "Sample sizes and all other conventions are the same as in Figure 5F.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/sivroni2023-auditory.html",
      "source_body_sha256": null,
      "evidence_sha256": "52065b9fee836c144127a0fbd0cb597a38102e1af5eb7094b3217a5b36250160",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "variance",
     "evidence": "During the example ses- sion, response times during CPTs and IPTs did not exhibit distinct medians (p = 0.15, U-test) or distinct variances (p = 0.49, permutation test).",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/sivroni2023-auditory.html",
      "source_body_sha256": null,
      "evidence_sha256": "c220d8a27529f6c3b00d817c82c28cb49c4cdfefae0e7bdfa1468be57d21ccbe",
      "grounding_status": "source-unavailable"
     }
    }
   ],
   "license_spdx": "personal-copy",
   "source_url": "https://scholar.google.com/scholar?q=Short-Term%20Auditory%20Priming%20in%20Freely-Moving%20Mice",
   "doi": "",
   "full_text_redistributable": false
  },
  {
   "id": "he2015-resnet",
   "title": "Deep Residual Learning for Image Recognition",
   "discipline": "artificial-intelligence",
   "year": 2015,
   "authors": "Kaiming He, Xiangyu Zhang, Shaoqing Ren, Jian Sun",
   "url": "papers/he2015-resnet.html",
   "key_terms": [
    "residual",
    "networks",
    "layers",
    "depth",
    "ImageNet",
    "identity",
    "shortcut",
    "accuracy",
    "error"
   ],
   "phenomenon": "training very deep neural networks (deep residual learning)",
   "claim_type": "method",
   "method": "computational-model",
   "statistics": [
    {
     "term": "mean",
     "evidence": "For ResNet-110, we run it 5 times and show “best (mean±std)” as in [43].",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "numeric mean or average in main text",
     "provenance": {
      "source_path": "papers/he2015-resnet.html",
      "source_body_sha256": "54dff8adcf3e9f2ec92692b24fd9ff624562f713b532416ee90b20d4dd4b3c94",
      "evidence_sha256": "1eff79cc6e53ad2dce46779c7106ff33effd831177875112db00782feeb72921",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "regression",
     "evidence": "Each dimension is binary logistic regression for predicting being or not be- ing an object class.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/he2015-resnet.html",
      "source_body_sha256": "54dff8adcf3e9f2ec92692b24fd9ff624562f713b532416ee90b20d4dd4b3c94",
      "evidence_sha256": "99491af4204e9804bac183b3fc53ba1cd8bb1ed213288358560270827e804fe1",
      "grounding_status": "not-exactly-located"
     }
    },
    {
     "term": "standard-deviation",
     "evidence": "Standard deviations (std) of layer responses on CIFAR- 10.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/he2015-resnet.html",
      "source_body_sha256": "54dff8adcf3e9f2ec92692b24fd9ff624562f713b532416ee90b20d4dd4b3c94",
      "evidence_sha256": "4716bfa96279983049eb55b180ad1dfc3826778174419f9f4b1e9c601d4f0d82",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "statistical-significance",
     "evidence": "Remarkably, although the depth is significantly increased, the 152-layer ResNet (11.3 billion FLOPs) still has lower complexity than VGG-16/19 nets (15.3/19.6 bil- lion FLOPs).",
     "decision": "rejected",
     "role": "lexical-significance",
     "reason": "significant/significantly is used without inferential-statistical meaning",
     "provenance": {
      "source_path": "papers/he2015-resnet.html",
      "source_body_sha256": "54dff8adcf3e9f2ec92692b24fd9ff624562f713b532416ee90b20d4dd4b3c94",
      "evidence_sha256": "bc99ce4c9e9f5587c2a759e1c7b3b4011c7d75f1e0f4ceff230b83e11bf64e42",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "arXiv-nonexclusive-distrib-1.0",
   "license_url": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
   "source_url": "https://arxiv.org/abs/1512.03385",
   "doi": "",
   "full_text_redistributable": false
  },
  {
   "id": "devlin2018-bert",
   "title": "BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding",
   "discipline": "artificial-intelligence",
   "year": 2018,
   "authors": "Jacob Devlin, Ming-Wei Chang, Kenton Lee, Kristina Toutanova",
   "url": "papers/devlin2018-bert.html",
   "key_terms": [
    "BERT",
    "pre-training",
    "fine-tuning",
    "bidirectional",
    "masked language model",
    "Transformer",
    "language representation",
    "next sentence prediction",
    "word embeddings",
    "self-attention",
    "natural language processing"
   ],
   "phenomenon": "pre-trained bidirectional language representations (BERT; contextual embeddings via Transformers)",
   "claim_type": "method",
   "method": "computational-model",
   "statistics": [
    {
     "term": "correlation",
     "evidence": "F1 scores are reported for QQP and MRPC, Spearman correlations are reported for STS-B, and accuracy scores are reported for the other tasks.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/devlin2018-bert.html",
      "source_body_sha256": "f6f5d06cd9d89748e7feae2e4a297d37773614b2bccdad2161578294a622b61a",
      "evidence_sha256": "31cc8d5b9708859899d7acb6e287f3734a9b8c4aaaa24d0381d7ec25b82b123a",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "arXiv-nonexclusive-distrib-1.0",
   "license_url": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
   "source_url": "https://arxiv.org/abs/1810.04805",
   "doi": "",
   "full_text_redistributable": false
  },
  {
   "id": "kingma2014-adam",
   "title": "Adam: A Method for Stochastic Optimization",
   "discipline": "artificial-intelligence",
   "year": 2014,
   "authors": "Diederik P. Kingma, Jimmy Ba",
   "url": "papers/kingma2014-adam.html",
   "key_terms": [],
   "phenomenon": "",
   "claim_type": "",
   "method": "computational-model",
   "statistics": [
    {
     "term": "Bayesian",
     "evidence": "e with high probability within some set region in parameter space; it is not uncommon, for example, to have a prior distribution over the parameters.",
     "decision": "rejected",
     "role": "insufficient-evidence",
     "reason": "supplied evidence does not establish the controlled term in the article body",
     "provenance": {
      "source_path": "papers/kingma2014-adam.html",
      "source_body_sha256": null,
      "evidence_sha256": "80469bede1afb9c95e34f412c7607984c3150b013171f40d750ab1293e5df28a",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "mean",
     "evidence": "Let g be the gradient of the stochastic objective f, and we wish to estimate its second raw moment (uncentered variance) using an exponential moving average of the squared gradient",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "numeric mean or average in main text",
     "provenance": {
      "source_path": "papers/kingma2014-adam.html",
      "source_body_sha256": null,
      "evidence_sha256": "092d009371208872f2f1fe4bd6de7a277aae33e4cdfeee42bea8ef03882388f6",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "regression",
     "evidence": "6 E XPERIMENTS To empirically evaluate the proposed method, we investigated different popular machine learning models, including logistic regression, multilayer fully connected neu",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/kingma2014-adam.html",
      "source_body_sha256": null,
      "evidence_sha256": "dd06bbbabeeef3e34d5168cc222a136b2155a7c054307d2c5c2514c131586a1d",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "variance",
     "evidence": "The moving averages themselves are estimates of the 1 st moment (the mean) and the 2nd raw moment (the uncentered variance) of the gradient.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/kingma2014-adam.html",
      "source_body_sha256": null,
      "evidence_sha256": "b69b14489e4bfdf05bf90c0b596b43cfc2e6f3112360fa15aef944b76c4581ba",
      "grounding_status": "source-unavailable"
     }
    }
   ],
   "license_spdx": "arxiv-open",
   "source_url": "https://arxiv.org/abs/1412.6980",
   "doi": "",
   "full_text_redistributable": false
  },
  {
   "id": "simonyan2014-vgg",
   "title": "Very Deep Convolutional Networks for Large-Scale Image Recognition",
   "discipline": "artificial-intelligence",
   "year": 2014,
   "authors": "Karen Simonyan, Andrew Zisserman",
   "url": "papers/simonyan2014-vgg.html",
   "key_terms": [
    "convolutional",
    "depth",
    "filters",
    "layers",
    "ImageNet",
    "classification",
    "network",
    "recognition",
    "localisation"
   ],
   "phenomenon": "very deep convolutional networks (image recognition; depth with small 3x3 filters)",
   "claim_type": "method",
   "method": "computational-model",
   "statistics": [
    {
     "term": "Bayesian",
     "evidence": "Learning generative visual models from few training examples: An In IEEE CVPR Workshop of Generative incremental bayesian approach tested on 101 object categories.",
     "decision": "rejected",
     "role": "bibliography",
     "reason": "occurrence is confined to a citation/reference rather than the article body",
     "provenance": {
      "source_path": "papers/simonyan2014-vgg.html",
      "source_body_sha256": "1aefb49f380080910fd821c842601e1d546ebba994a0ee3ff2816cef38dbbef3",
      "evidence_sha256": "1f616ed20919d3be2d650b21647f03f4d6e4895db2d400c0eb8a3ec624e90e59",
      "grounding_status": "verified-reference"
     }
    },
    {
     "term": "regression",
     "evidence": "Namely, the training is carried out by optimising the multinomial logistic regression objective using mini-batch gradient descent (based on back-propagation (LeCun et al., 1989)) w",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/simonyan2014-vgg.html",
      "source_body_sha256": "1aefb49f380080910fd821c842601e1d546ebba994a0ee3ff2816cef38dbbef3",
      "evidence_sha256": "c22be2f2ccb71385548c4c696a4d04d560a86b312984aba43e5e32b976c6cc69",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "statistical-significance",
     "evidence": "As a result, we come up with significantly more accurate ConvNet architectures, which not only achieve the state-of-the-art accuracy on ILSVRC classification and localisation tasks",
     "decision": "rejected",
     "role": "lexical-significance",
     "reason": "significant/significantly is used without inferential-statistical meaning",
     "provenance": {
      "source_path": "papers/simonyan2014-vgg.html",
      "source_body_sha256": "1aefb49f380080910fd821c842601e1d546ebba994a0ee3ff2816cef38dbbef3",
      "evidence_sha256": "c1b4fe4d82327ccb2d14d2cd100c82c7f3d8081617026124447bd304ed0541cb",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "arXiv-nonexclusive-distrib-1.0",
   "license_url": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
   "source_url": "https://arxiv.org/abs/1409.1556",
   "doi": "",
   "full_text_redistributable": false
  },
  {
   "id": "ronneberger2015-unet",
   "title": "U-Net: Convolutional Networks for Biomedical Image Segmentation",
   "discipline": "medicine",
   "year": 2015,
   "authors": "Olaf Ronneberger, Philipp Fischer, Thomas Brox",
   "url": "papers/ronneberger2015-unet.html",
   "key_terms": [
    "network",
    "segmentation",
    "images",
    "training",
    "convolutional",
    "localization",
    "augmentation",
    "pixel",
    "architecture"
   ],
   "phenomenon": "biomedical image segmentation (U-Net; deep learning from few images)",
   "claim_type": "method",
   "method": "computational-model",
   "statistics": [
    {
     "term": "standard-deviation",
     "evidence": "tion and ReLU layers) this can be achieved by drawing the initial weights from a Gaussian distribution with a standard deviation of (cid:112)2/N , where N denotes the number of inc",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/ronneberger2015-unet.html",
      "source_body_sha256": "daccf9478244ba80b12bf19ef7893e3036518fc83e13bdc9368945b5207882d6",
      "evidence_sha256": "01397442f8cd04ecb7c7e5603361b81afa34105cbd45e4ef764ae563bf4291aa",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "standard-error",
     "evidence": "Name PhC-U373 DIC-HeLa IMCB-SG (2014) KTH-SE (2014) HOUS-US (2014) second-best 2015 u-net (2015) 0.2669 0.7953 0.5323 0.83 0.9203 0.2935 0.4607 - 0.46 0.7",
     "decision": "rejected",
     "role": "non-statistical-sense",
     "reason": "lexical match does not denote the controlled statistical concept",
     "provenance": {
      "source_path": "papers/ronneberger2015-unet.html",
      "source_body_sha256": "daccf9478244ba80b12bf19ef7893e3036518fc83e13bdc9368945b5207882d6",
      "evidence_sha256": "dd1ffcd84972acc7abbd8ec838acf16853427f8e123f5f677409ffd97f7ee281",
      "grounding_status": "not-exactly-located"
     }
    }
   ],
   "license_spdx": "arXiv-nonexclusive-distrib-1.0",
   "license_url": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
   "source_url": "https://arxiv.org/abs/1505.04597",
   "doi": "",
   "full_text_redistributable": false
  },
  {
   "id": "srivastava2014-dropout",
   "title": "Dropout: A Simple Way to Prevent Neural Networks from Overfitting",
   "discipline": "artificial-intelligence",
   "year": 2014,
   "authors": "Nitish Srivastava, Geoffrey Hinton, Alex Krizhevsky, Ilya Sutskever, Ruslan Salakhutdinov",
   "url": "papers/srivastava2014-dropout.html",
   "key_terms": [
    "dropout",
    "overfitting",
    "neural networks",
    "regularization",
    "thinned networks",
    "model combination",
    "co-adapting",
    "deep learning",
    "training",
    "hidden units"
   ],
   "phenomenon": "regularization of neural networks (preventing overfitting)",
   "claim_type": "method",
   "method": "computational-model",
   "statistics": [
    {
     "term": "Bayesian",
     "evidence": "This can sometimes be approximated quite well for simple or small models (Xiong et al., 2011; Salakhutdinov and Mnih, 2008), but we would like to approach the performance of the Ba",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/srivastava2014-dropout.html",
      "source_body_sha256": "7814371c6462e4c2f648c193303a8c6a427fca9b86ae03ff5fd3286dc25c17e4",
      "evidence_sha256": "10ace4f6a4d7f39216a8deb4e81739e060899058e62b93858ed7402a0e3a83ae",
      "grounding_status": "not-exactly-located"
     }
    },
    {
     "term": "mean",
     "evidence": "We propose to do this by approximating an equally weighted geometric mean of the predictions of an exponential number of learned models that share parameters.",
     "decision": "rejected",
     "role": "non-statistical-sense",
     "reason": "lexical match does not denote the controlled statistical concept",
     "provenance": {
      "source_path": "papers/srivastava2014-dropout.html",
      "source_body_sha256": "7814371c6462e4c2f648c193303a8c6a427fca9b86ae03ff5fd3286dc25c17e4",
      "evidence_sha256": "d57bdb380b8ca004c52c0d43378e7ce64c43db141dca208df6cd13085ee2835c",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "p-value",
     "evidence": "All dropout nets use p = 0.5 for hidden units and p = 0.8 for input units.",
     "decision": "rejected",
     "role": "non-statistical-sense",
     "reason": "lexical match does not denote the controlled statistical concept",
     "provenance": {
      "source_path": "papers/srivastava2014-dropout.html",
      "source_body_sha256": "7814371c6462e4c2f648c193303a8c6a427fca9b86ae03ff5fd3286dc25c17e4",
      "evidence_sha256": "971dd31bf295e4276850cdec2f936a15afec5c2d3bc16404ca8fb6b6cb6dadbd",
      "grounding_status": "not-exactly-located"
     }
    },
    {
     "term": "regression",
     "evidence": "The marginal- ization of dropout noise in the context of linear regression was discussed in Srivastava (2013).",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/srivastava2014-dropout.html",
      "source_body_sha256": "7814371c6462e4c2f648c193303a8c6a427fca9b86ae03ff5fd3286dc25c17e4",
      "evidence_sha256": "5b70b46bb4b9fabfcce40b5819d1ac053bc572180a278ec186117fc73cd5b65e",
      "grounding_status": "not-exactly-located"
     }
    },
    {
     "term": "standard-deviation",
     "evidence": "Thereafter, the Monte-Carlo method is slightly better than the approximate method but well within one standard deviation of it.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/srivastava2014-dropout.html",
      "source_body_sha256": "7814371c6462e4c2f648c193303a8c6a427fca9b86ae03ff5fd3286dc25c17e4",
      "evidence_sha256": "16c0a34c88c769831e8b19e654b2a15edd6ce4f209b81c1d7a159315cccf4900",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "statistical-significance",
     "evidence": "found that training a network with dropout and using this approximate averaging method at test time leads to significantly lower generalization error on a wide variety of classific",
     "decision": "rejected",
     "role": "lexical-significance",
     "reason": "significant/significantly is used without inferential-statistical meaning",
     "provenance": {
      "source_path": "papers/srivastava2014-dropout.html",
      "source_body_sha256": "7814371c6462e4c2f648c193303a8c6a427fca9b86ae03ff5fd3286dc25c17e4",
      "evidence_sha256": "10243bf993b5f3758068c1c9d54943c7a2547e59fbbbb6232f065b7db8ab9784",
      "grounding_status": "not-exactly-located"
     }
    }
   ],
   "license_spdx": "arxiv-open",
   "license_url": "",
   "source_url": "https://jmlr.org/papers/v15/srivastava14a.html",
   "doi": "",
   "full_text_redistributable": false
  },
  {
   "id": "atlas2012-higgs",
   "title": "Observation of a New Particle in the Search for the Standard Model Higgs Boson with the ATLAS Detector",
   "discipline": "physics",
   "year": 2012,
   "authors": "ATLAS Collaboration",
   "url": "papers/atlas2012-higgs.html",
   "key_terms": [
    "Higgs boson",
    "Standard Model",
    "ATLAS",
    "LHC",
    "electroweak symmetry",
    "proton-proton collisions",
    "significance",
    "standard deviations",
    "luminosity",
    "scalar particle"
   ],
   "phenomenon": "Higgs boson (discovery of a new particle at 5-sigma significance)",
   "claim_type": "existence",
   "method": "experiment",
   "statistics": [
    {
     "term": "chi-square",
     "evidence": ") WW ( ∗ γγ and H Asymptotically, the test statistic 2 ln λ(μ, mH) is dis- tributed as a χ2 distribution with two degrees of free- dom.",
     "decision": "accepted",
     "role": "substantive-mention",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/atlas2012-higgs.html",
      "source_body_sha256": "ed2f6700bfd4a907eb5c20a8a3334c8e06655ad113cb164b6c031a4a93f2a2a7",
      "evidence_sha256": "67c3c1775537d6b78c73fb2fb306e87ee7e61f337c1a88144a490f3add644293",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "confidence-interval",
     "evidence": "11, although they are only approximate confidence intervals due to the smaller number of can- didates in this channel.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/atlas2012-higgs.html",
      "source_body_sha256": "ed2f6700bfd4a907eb5c20a8a3334c8e06655ad113cb164b6c031a4a93f2a2a7",
      "evidence_sha256": "fe1819e05477ac7124ee0f61600ba451f65009a41a1fb75b75f9ca772c5c50d3",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "p-value",
     "evidence": "The horizontal dashed lines indicate the p-values corresponding to significances of 1 to 6 σ.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit p-value or inferential p notation",
     "provenance": {
      "source_path": "papers/atlas2012-higgs.html",
      "source_body_sha256": "ed2f6700bfd4a907eb5c20a8a3334c8e06655ad113cb164b6c031a4a93f2a2a7",
      "evidence_sha256": "df4b742ec9d2c94fd4230e011c65140c67a70f2b110749b96a121ef33237b601",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "standard-deviation",
     "evidence": "production of a neutral boson with a measured mass of 126.0 This observation, which has a significance of 5.9 standard deviations, corresponding to a background 9, is compatible wi",
     "decision": "accepted",
     "role": "substantive-mention",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/atlas2012-higgs.html",
      "source_body_sha256": "ed2f6700bfd4a907eb5c20a8a3334c8e06655ad113cb164b6c031a4a93f2a2a7",
      "evidence_sha256": "655ab130a9453530a73106bc54fb11c02623dd6531aa8ad96828997bd2b079de",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "statistical-power",
     "evidence": "Different models are chosen for the different categories to achieve a good compromise between limiting the size of a potential bias while retaining good statistical power.",
     "decision": "accepted",
     "role": "substantive-mention",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/atlas2012-higgs.html",
      "source_body_sha256": "ed2f6700bfd4a907eb5c20a8a3334c8e06655ad113cb164b6c031a4a93f2a2a7",
      "evidence_sha256": "46a96aa74e7dc4980d1e443c72c4186f3fb1491119271cd06906208184909d91",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "statistical-significance",
     "evidence": "The horizontal dashed lines indicate the p-values corresponding to significances of 1 to 6 σ.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/atlas2012-higgs.html",
      "source_body_sha256": "ed2f6700bfd4a907eb5c20a8a3334c8e06655ad113cb164b6c031a4a93f2a2a7",
      "evidence_sha256": "df4b742ec9d2c94fd4230e011c65140c67a70f2b110749b96a121ef33237b601",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "arXiv-nonexclusive-distrib-1.0",
   "license_url": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
   "source_url": "https://arxiv.org/abs/1207.7214",
   "doi": "",
   "full_text_redistributable": false
  },
  {
   "id": "maldacena1997-adscft",
   "title": "The Large N Limit of Superconformal Field Theories and Supergravity",
   "discipline": "physics",
   "year": 1998,
   "authors": "Juan M. Maldacena",
   "url": "papers/maldacena1997-adscft.html",
   "key_terms": [
    "conformal field theories",
    "supergravity",
    "large N limit",
    "Anti-deSitter",
    "super-Yang-Mills",
    "branes",
    "superconformal group",
    "string theory",
    "M-theory",
    "near horizon geometry",
    "duality"
   ],
   "phenomenon": "gauge/gravity duality (AdS/CFT correspondence; holography)",
   "claim_type": "prediction",
   "method": "theory",
   "statistics": [],
   "license_spdx": "arxiv-open",
   "license_url": "",
   "source_url": "https://arxiv.org/abs/hep-th/9711200",
   "doi": "10.4310/ATMP.1998.v2.n2.a1",
   "full_text_redistributable": false
  },
  {
   "id": "abbott2016-gravitational-waves",
   "title": "Observation of Gravitational Waves from a Binary Black Hole Merger",
   "discipline": "astronomy",
   "year": 2016,
   "authors": "B. P. Abbott et al. (LIGO Scientific Collaboration and Virgo Collaboration)",
   "url": "papers/abbott2016-gravitational-waves.html",
   "key_terms": [
    "gravitational",
    "black hole",
    "merger",
    "LIGO",
    "strain",
    "signal",
    "relativity",
    "waveform",
    "detectors"
   ],
   "phenomenon": "gravitational waves (first direct detection; binary black hole merger)",
   "claim_type": "existence",
   "method": "experiment",
   "statistics": [
    {
     "term": "Bayesian",
     "evidence": "To refine them we use general relativity-based models [77,78,87,88], some of which include spin precession, and for each model perform a coherent Bayesian analysis to derive poster",
     "decision": "accepted",
     "role": "substantive-mention",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/abbott2016-gravitational-waves.html",
      "source_body_sha256": "fca20fe44927978fd8d30a9cec96a0c33dfa01bd57775411fef4878b7e4f49cf",
      "evidence_sha256": "37cc9c521b79dd051a85c50068a50980585c87dd37b276e5ff02edfc37714494",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "chi-square",
     "evidence": "For each maximum we calculate a chi-squared statistic χ2 to test whether the data in several different r frequency bands are consistent with the matching template [82].",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/abbott2016-gravitational-waves.html",
      "source_body_sha256": "fca20fe44927978fd8d30a9cec96a0c33dfa01bd57775411fef4878b7e4f49cf",
      "evidence_sha256": "c20370c33a5075daac4990eff6109047ddb5387152d6c04d5612381eae3331e0",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "median",
     "evidence": "We report median values with 90% credible intervals that include statistical errors, and systematic errors from averaging the results of different waveform models.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/abbott2016-gravitational-waves.html",
      "source_body_sha256": "fca20fe44927978fd8d30a9cec96a0c33dfa01bd57775411fef4878b7e4f49cf",
      "evidence_sha256": "64f4e26c8497557a8be58bf228fa1b12e4cec3de88e01d7173029b417adc825b",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "standard-deviation",
     "evidence": "The scales on the top give the significance of an event in Gaussian standard deviations based on the corresponding noise background.",
     "decision": "accepted",
     "role": "substantive-mention",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/abbott2016-gravitational-waves.html",
      "source_body_sha256": "fca20fe44927978fd8d30a9cec96a0c33dfa01bd57775411fef4878b7e4f49cf",
      "evidence_sha256": "92d218139e35c009932e8cfc35c6a1e99e1a35712b1928293d29e0dda9efab6b",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "statistical-significance",
     "evidence": "noise ratio of 24 and a false alarm rate estimated to be less than 1 event per 203 000 years, equivalent to a significance greater −180 Mpc corresponding to a redshift z 1⁄4 0.09þ0",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/abbott2016-gravitational-waves.html",
      "source_body_sha256": "fca20fe44927978fd8d30a9cec96a0c33dfa01bd57775411fef4878b7e4f49cf",
      "evidence_sha256": "3c4f878cc5c7d668f2e5b1032c3411474bf5537b1d30cdc9c89e6852bb7c2da8",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "CC-BY-4.0",
   "license_url": "http://creativecommons.org/licenses/by/4.0/",
   "source_url": "https://arxiv.org/abs/1602.03837",
   "doi": "10.1103/PhysRevLett.116.061102",
   "full_text_redistributable": true
  },
  {
   "id": "novoselov2004-graphene",
   "title": "Electric Field Effect in Atomically Thin Carbon Films",
   "discipline": "materials-science",
   "year": 2004,
   "authors": "K. S. Novoselov, A. K. Geim, et al.",
   "url": "papers/novoselov2004-graphene.html",
   "key_terms": [
    "graphene",
    "carbon",
    "two-dimensional",
    "electrons",
    "holes",
    "semimetal",
    "mobility",
    "field effect",
    "atoms"
   ],
   "phenomenon": "graphene (stable two-dimensional atomically-thin carbon; electric field effect)",
   "claim_type": "existence",
   "method": "experiment",
   "statistics": [],
   "license_spdx": "arxiv-open",
   "license_url": "",
   "source_url": "https://arxiv.org/abs/cond-mat/0410550",
   "doi": "10.1126/science.1102896",
   "full_text_redistributable": false
  },
  {
   "id": "shor1997-factoring",
   "title": "Polynomial-Time Algorithms for Prime Factorization and Discrete Logarithms on a Quantum Computer",
   "discipline": "quantum-computing",
   "year": 1997,
   "authors": "Peter W. Shor",
   "url": "papers/shor1997-factoring.html",
   "key_terms": [
    "factoring",
    "quantum",
    "algorithm",
    "computer",
    "integer",
    "polynomial",
    "logarithms",
    "prime",
    "factorization"
   ],
   "phenomenon": "quantum algorithms for factoring and discrete logarithm (exponential speedup)",
   "claim_type": "method",
   "method": "mathematical-proof",
   "statistics": [],
   "license_spdx": "arxiv-open",
   "license_url": "",
   "source_url": "https://arxiv.org/abs/quant-ph/9508027",
   "doi": "10.1137/S0097539795293172",
   "full_text_redistributable": false
  },
  {
   "id": "perelman2002-ricci-flow",
   "title": "The Entropy Formula for the Ricci Flow and Its Geometric Applications",
   "discipline": "mathematics",
   "year": 2002,
   "authors": "Grigori Perelman",
   "url": "papers/perelman2002-ricci-flow.html",
   "key_terms": [
    "Ricci flow",
    "curvature",
    "metric",
    "manifold",
    "entropy",
    "monotonicity",
    "geometrization",
    "surgery",
    "soliton"
   ],
   "phenomenon": "Ricci flow and the geometrization/Poincare conjecture (entropy monotonicity)",
   "claim_type": "method",
   "method": "mathematical-proof",
   "statistics": [],
   "license_spdx": "arxiv-open",
   "license_url": "",
   "source_url": "https://arxiv.org/abs/math/0211159",
   "doi": "",
   "full_text_redistributable": false
  },
  {
   "id": "newman2003-complex-networks",
   "title": "The Structure and Function of Complex Networks",
   "discipline": "network-science",
   "year": 2003,
   "authors": "M. E. J. Newman",
   "url": "papers/newman2003-complex-networks.html",
   "key_terms": [
    "networks",
    "degree",
    "clustering",
    "vertices",
    "edges",
    "scale-free",
    "small-world",
    "distribution",
    "preferential"
   ],
   "phenomenon": "complex networks (shared structure and function of real-world networks)",
   "claim_type": "existence",
   "method": "theory",
   "statistics": [
    {
     "term": "correlation",
     "evidence": "For the data of Table III we find r = 0.621.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/newman2003-complex-networks.html",
      "source_body_sha256": "eb63d7c9c83fd176155b1e518ddf4c80ec0f91c707ed92f25c7583a94967e384",
      "evidence_sha256": "8dbb12f0dcb0f002bed80ca94772b89a5bda8911beb65235f92ef90bf179e735",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "hypothesis-testing",
     "evidence": "l graphs which have similar statistical properties to their real-world counterparts and which can be used for hypothesis testing or as a substrate for further network simulations.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit hypothesis-testing language",
     "provenance": {
      "source_path": "papers/newman2003-complex-networks.html",
      "source_body_sha256": "eb63d7c9c83fd176155b1e518ddf4c80ec0f91c707ed92f25c7583a94967e384",
      "evidence_sha256": "a1570d64b968bb7ee1306bfd781d0e08a1fc58f75c1aa91b7e32b58d556ab33c",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "mean",
     "evidence": "(5), of 1, 1, 1 6 , 0 and 0, for a mean value, Eq.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/newman2003-complex-networks.html",
      "source_body_sha256": "eb63d7c9c83fd176155b1e518ddf4c80ec0f91c707ed92f25c7583a94967e384",
      "evidence_sha256": "c6921c6792d96d10203ca6143d43331b1ed53d4b9ecded25810b575be488706d",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "null-hypothesis",
     "evidence": "n of the size of the component of occupied vertices to which a randomly chosen vertex belongs is generated by H0(x) where 0.03 0.02 0.01 n o i t c a r f l a c i t i r c H0(x) = 1 H",
     "decision": "rejected",
     "role": "symbol-collision",
     "reason": "H0 denotes a Hubble parameter, Hamiltonian, operator, or generating function rather than a null hypothesis",
     "provenance": {
      "source_path": "papers/newman2003-complex-networks.html",
      "source_body_sha256": "eb63d7c9c83fd176155b1e518ddf4c80ec0f91c707ed92f25c7583a94967e384",
      "evidence_sha256": "a08b819354ad44334c65b8f2eba44bf9ac6d6ac1e1719b91df4908b90ab3d548",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "regression",
     "evidence": "and Pattison, P., Logit models and lo- gistic regressions for social networks: I.",
     "decision": "rejected",
     "role": "insufficient-evidence",
     "reason": "supplied evidence does not establish the controlled term in the article body",
     "provenance": {
      "source_path": "papers/newman2003-complex-networks.html",
      "source_body_sha256": "eb63d7c9c83fd176155b1e518ddf4c80ec0f91c707ed92f25c7583a94967e384",
      "evidence_sha256": "790b59d265cb311d36ebe0c2c24c68a9f1c81bf9fea7f362f771b12e082372b5",
      "grounding_status": "verified-reference"
     }
    },
    {
     "term": "sample-size",
     "evidence": "Traditional social network studies often suffer from problems of inaccuracy, subjectivity, and small sample size.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/newman2003-complex-networks.html",
      "source_body_sha256": "eb63d7c9c83fd176155b1e518ddf4c80ec0f91c707ed92f25c7583a94967e384",
      "evidence_sha256": "8ce4fadd4d77505752b49578500a05947183410908fcba4595e176178e4c0243",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "statistical-significance",
     "evidence": "This assumption however degree distribution pk ∼ can give misleading results; in many cases there will be vertices in the network with significantly higher degree than this, as dis",
     "decision": "rejected",
     "role": "lexical-significance",
     "reason": "significant/significantly is used without inferential-statistical meaning",
     "provenance": {
      "source_path": "papers/newman2003-complex-networks.html",
      "source_body_sha256": "eb63d7c9c83fd176155b1e518ddf4c80ec0f91c707ed92f25c7583a94967e384",
      "evidence_sha256": "98dcec2bed607e08495245d508eca6c1f88210daa4cd17d1bc82492622b46a1d",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "arxiv-open",
   "license_url": "",
   "source_url": "https://arxiv.org/abs/cond-mat/0303516",
   "doi": "10.1137/S003614450342480",
   "full_text_redistributable": false
  },
  {
   "id": "nakamoto2008-bitcoin",
   "title": "Bitcoin: A Peer-to-Peer Electronic Cash System",
   "discipline": "economics",
   "year": 2008,
   "authors": "Satoshi Nakamoto",
   "url": "papers/nakamoto2008-bitcoin.html",
   "key_terms": [
    "network",
    "proof-of-work",
    "hash",
    "chain",
    "transactions",
    "nodes",
    "block",
    "timestamp",
    "peer-to-peer"
   ],
   "phenomenon": "decentralized trustless digital currency (distributed consensus without a trusted third party)",
   "claim_type": "method",
   "method": "theory",
   "statistics": [
    {
     "term": "p-value",
     "evidence": "q=0.1 z=0 P=1.0000000 z=1 P=0.2045873 z=2 P=0.0509779 z=3 P=0.0131722 z=4 P=0.0034552 z=5 P=0.0009137 z=6 P=0.0002428 z=7 P=0.0000647 z=8 P=0.0000173 z=9 P=0.0000046 z=10 P=0.00000",
     "decision": "rejected",
     "role": "non-statistical-sense",
     "reason": "lexical match does not denote the controlled statistical concept",
     "provenance": {
      "source_path": "papers/nakamoto2008-bitcoin.html",
      "source_body_sha256": "353c07c40e9c733a7a48d17bdd8aa93d4b3f8703aac8f863301efda64951a340",
      "evidence_sha256": "43c45a1c2402d23006a7f7bbdedd970cc9a215638a097a86c45a9cd24b1d197e",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "open-access",
   "license_url": "",
   "source_url": "https://bitcoin.org/bitcoin.pdf",
   "doi": "",
   "full_text_redistributable": false
  },
  {
   "id": "riess1998-accelerating-universe",
   "title": "Observational Evidence from Supernovae for an Accelerating Universe and a Cosmological Constant",
   "discipline": "astronomy",
   "year": 1998,
   "authors": "Adam G. Riess et al. (High-z Supernova Search Team)",
   "url": "papers/riess1998-accelerating-universe.html",
   "key_terms": [
    "supernovae",
    "redshift",
    "luminosity",
    "distance",
    "cosmological",
    "universe",
    "expansion",
    "acceleration"
   ],
   "phenomenon": "accelerating expansion of the universe (dark energy / cosmological constant)",
   "claim_type": "existence",
   "method": "observation",
   "statistics": [
    {
     "term": "Bayesian",
     "evidence": "These extinction measurements employ the same Bayesian filter (in the Appendix) used for the MLCS fits.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/riess1998-accelerating-universe.html",
      "source_body_sha256": "60a9ce4221fc7fbd9795e34410623af9ad555924bcf6a3ff02da27c91f70f582",
      "evidence_sha256": "81ba41545175131b7edaa974a7f2302d42a44dde35df8e7210a222b86d136a67",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "chi-square",
     "evidence": "The likelihood for the cosmological parameters can be determined from a χ2 statistic, where χ2(H0, ΩM , ΩΛ) = i X (μp,i(zi; H0, Ωm, ΩΛ) − μ0,i)2 μ0,i + σ2 σ2 v (4) – 12 – and σv is",
     "decision": "accepted",
     "role": "substantive-mention",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/riess1998-accelerating-universe.html",
      "source_body_sha256": "60a9ce4221fc7fbd9795e34410623af9ad555924bcf6a3ff02da27c91f70f582",
      "evidence_sha256": "571119b70aac7665e0c431bec574a5d6b96aa44a58186f3a09e03a4b6dcc0821",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "confidence-interval",
     "evidence": "The χ2 statistic of equation (4) is well suited for determining the most likely values for the cosmological parameters H0, ΩM , and ΩΛ as well as the confidence intervals surroundi",
     "decision": "accepted",
     "role": "substantive-mention",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/riess1998-accelerating-universe.html",
      "source_body_sha256": "60a9ce4221fc7fbd9795e34410623af9ad555924bcf6a3ff02da27c91f70f582",
      "evidence_sha256": "45ef7d7082d4c16cf3e1a7e11b17677813128d864b03c5f711eb77e51f51c3c8",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "correlation",
     "evidence": "nce in maximum luminosity between the fiducial template SN Ia and any other SN Ia, RV and RB−V are vectors of correlation coefficients between ∆ and the light curve shape, μv is th",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/riess1998-accelerating-universe.html",
      "source_body_sha256": "60a9ce4221fc7fbd9795e34410623af9ad555924bcf6a3ff02da27c91f70f582",
      "evidence_sha256": "098c49ef90a9fe1b57f4ae22b3617b5710936d7072226566f157a42b07adbbd2",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "mean",
     "evidence": "Yet the dispersion of individual extinctions for a distribution with a mean of 0.25 mag would be σAB =0.40 mag, significantly larger than the 0.21 mag dispersion observed in the hi",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/riess1998-accelerating-universe.html",
      "source_body_sha256": "60a9ce4221fc7fbd9795e34410623af9ad555924bcf6a3ff02da27c91f70f582",
      "evidence_sha256": "2fe28892451435e3c9ce89e919324cec8e7174e693f55ce48b9c54eeff08751c",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "null-hypothesis",
     "evidence": "The likelihood for the cosmological parameters can be determined from a χ2 statistic, where χ2(H0, ΩM , ΩΛ) = i X (μp,i(zi; H0, Ωm, ΩΛ) − μ0,i)2 μ0,i + σ2 σ2 v (4) – 12 – and σv is",
     "decision": "rejected",
     "role": "symbol-collision",
     "reason": "H0 denotes a Hubble parameter, Hamiltonian, operator, or generating function rather than a null hypothesis",
     "provenance": {
      "source_path": "papers/riess1998-accelerating-universe.html",
      "source_body_sha256": "60a9ce4221fc7fbd9795e34410623af9ad555924bcf6a3ff02da27c91f70f582",
      "evidence_sha256": "571119b70aac7665e0c431bec574a5d6b96aa44a58186f3a09e03a4b6dcc0821",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "sample-size",
     "evidence": "Confidence in these results depends not on increasing the sample size but on improving our understanding of systematic uncertainties.",
     "decision": "accepted",
     "role": "substantive-mention",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/riess1998-accelerating-universe.html",
      "source_body_sha256": "60a9ce4221fc7fbd9795e34410623af9ad555924bcf6a3ff02da27c91f70f582",
      "evidence_sha256": "a7079aab1b1f418fe7466d34ffdb81354ec57fb387794ab0c185c4961b58ec43",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "statistical-significance",
     "evidence": "1998a), the sample of 16 is now large enough to yield interesting cosmological results of high statistical significance.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit inferential-statistical context",
     "provenance": {
      "source_path": "papers/riess1998-accelerating-universe.html",
      "source_body_sha256": "60a9ce4221fc7fbd9795e34410623af9ad555924bcf6a3ff02da27c91f70f582",
      "evidence_sha256": "d294182c7d4efb98b45c7347767598a91ac7773dd2ec7e46853c04d12c9304a9",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "variance",
     "evidence": "These determinations include only the statistical component of error resulting from the point-to-point variance of the measured Hubble flow and do not include any uncertainty in th",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/riess1998-accelerating-universe.html",
      "source_body_sha256": "60a9ce4221fc7fbd9795e34410623af9ad555924bcf6a3ff02da27c91f70f582",
      "evidence_sha256": "a74f05becac23d035192d2fe16ba4dda53ff795fbde28b002bf05dc4cfdb7562",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "arxiv-open",
   "license_url": "",
   "source_url": "https://arxiv.org/abs/astro-ph/9805201",
   "doi": "10.1086/300499",
   "full_text_redistributable": false
  },
  {
   "id": "vonelm2007-strobe",
   "title": "The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) Statement",
   "discipline": "medicine",
   "year": 2007,
   "authors": "Erik von Elm, Douglas G. Altman, Matthias Egger, et al.",
   "url": "papers/vonelm2007-strobe.html",
   "key_terms": [
    "reporting",
    "observational",
    "studies",
    "checklist",
    "STROBE",
    "recommendations",
    "epidemiology",
    "transparent",
    "items"
   ],
   "phenomenon": "transparent reporting standards for observational studies (reducing bias)",
   "claim_type": "method",
   "method": "survey",
   "statistics": [
    {
     "term": "confidence-interval",
     "evidence": "rt numbers of outcome events or summary measures (a) Give unadjusted estimates and, if applicable, confounder-adjusted estimates and their precision (e.g., 95% confidence interval)",
     "decision": "accepted",
     "role": "substantive-mention",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/vonelm2007-strobe.html",
      "source_body_sha256": "f691e16a47ed98b59f4e18c4a9a8e4b2300109a7c349f5739ecb4cd217af48be",
      "evidence_sha256": "eeb7f6f55478eb4f3a6876620fe5091a8e198f43adf72c620e4e07dbfc3e9ab9",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "odds-ratio",
     "evidence": "gory boundaries when continuous variables were categorized (c) If relevant, consider translating estimates of relative risk into absolute risk for a meaningful time period Report o",
     "decision": "rejected",
     "role": "non-statistical-sense",
     "reason": "lexical match does not denote the controlled statistical concept",
     "provenance": {
      "source_path": "papers/vonelm2007-strobe.html",
      "source_body_sha256": "f691e16a47ed98b59f4e18c4a9a8e4b2300109a7c349f5739ecb4cd217af48be",
      "evidence_sha256": "65f8c86ce082061ada215c248888549f09ec5b1b3388526338293455914dbcf0",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "open-access",
   "license_url": "",
   "source_url": "https://journals.plos.org/plosmedicine/article?id=10.1371/journal.pmed.0040296",
   "doi": "10.1371/journal.pmed.0040296",
   "full_text_redistributable": false
  },
  {
   "id": "lake2017-machines-learn-think",
   "title": "Building Machines That Learn and Think Like People",
   "discipline": "cognitive-science",
   "year": 2017,
   "authors": "Brenden M. Lake, Tomer D. Ullman, Joshua B. Tenenbaum, Samuel J. Gershman",
   "url": "papers/lake2017-machines-learn-think.html",
   "key_terms": [
    "artificial intelligence",
    "deep learning",
    "neural networks",
    "human intelligence",
    "causal models",
    "intuitive theories",
    "compositionality",
    "learning-to-learn",
    "cognitive science",
    "pattern recognition"
   ],
   "phenomenon": "human-like learning and cognition in AI (causal models, intuitive physics/psychology, compositionality)",
   "claim_type": "critique",
   "method": "theory",
   "statistics": [
    {
     "term": "Bayesian",
     "evidence": "We will generally be concerned with directed generative models (such as Bayesian networks or probabilistic programs) which can be given a causal interpretation, although undirected",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/lake2017-machines-learn-think.html",
      "source_body_sha256": "9aa7e33520db38a8f24b445c4c3a514c6074b22516ceba471762efb606785c30",
      "evidence_sha256": "b0cf68781affafd88db4e21060251888fd5518d2c41f55dba522002811413f99",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "mean",
     "evidence": "(2015) scored an average of 4335 points across 30 game sessions of up to five minutes of play.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/lake2017-machines-learn-think.html",
      "source_body_sha256": "9aa7e33520db38a8f24b445c4c3a514c6074b22516ceba471762efb606785c30",
      "evidence_sha256": "b32fe343ee2269500b45b70009b5333014360eedd2a0b3b513239ef8f0105f53",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "regression",
     "evidence": "The statistical pattern recognition approach treats prediction as primary, usually in the context of a specific classification, regression, or control task.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/lake2017-machines-learn-think.html",
      "source_body_sha256": "9aa7e33520db38a8f24b445c4c3a514c6074b22516ceba471762efb606785c30",
      "evidence_sha256": "ec7934b8dd761f7eabb8352ec59c107f4f33ad3090783d0bae71756fabcaa5a9",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "statistical-significance",
     "evidence": "We believe that integrating them will produce significantly more powerful and more human-like learning and thinking abilities than we currently see in AI systems.",
     "decision": "rejected",
     "role": "lexical-significance",
     "reason": "significant/significantly is used without inferential-statistical meaning",
     "provenance": {
      "source_path": "papers/lake2017-machines-learn-think.html",
      "source_body_sha256": "9aa7e33520db38a8f24b445c4c3a514c6074b22516ceba471762efb606785c30",
      "evidence_sha256": "799a678f1a6ca6ce89e6d0f7be4ac4ff7e15ffedc10ec8c905c91db08beb7605",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "arXiv-nonexclusive-distrib-1.0",
   "license_url": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
   "source_url": "https://arxiv.org/abs/1604.00289",
   "doi": "10.1017/S0140525X16001837",
   "full_text_redistributable": false
  },
  {
   "id": "jeong2000-metabolic-networks",
   "title": "The Large-Scale Organization of Metabolic Networks",
   "discipline": "biology",
   "year": 2000,
   "authors": "H. Jeong, B. Tombor, R. Albert, Z. N. Oltvai, A.-L. Barabási",
   "url": "papers/jeong2000-metabolic-networks.html",
   "key_terms": [
    "metabolic",
    "network",
    "substrate",
    "scale-free",
    "power-law",
    "connectivity",
    "diameter",
    "hubs",
    "reactions"
   ],
   "phenomenon": "scale-free network (power-law degree distribution in metabolic networks)",
   "claim_type": "existence",
   "method": "computational-model",
   "statistics": [
    {
     "term": "standard-deviation",
     "evidence": "To quantify this observation, we examined the standard deviation (s r) of the rank for substrates that are present in all 43 organisms.",
     "decision": "accepted",
     "role": "substantive-mention",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/jeong2000-metabolic-networks.html",
      "source_body_sha256": "80328c1d513de4e0a1c50f0f78236b6b1d0eb92e3a825b6f5aaa164d9c584c00",
      "evidence_sha256": "60338f123510bea1f7cc9b5b06f8b6afb7a872b9622c5078d807901aef6e7452",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "arxiv-open",
   "license_url": "",
   "source_url": "https://arxiv.org/abs/cond-mat/0010278",
   "doi": "10.1038/35036627",
   "full_text_redistributable": false
  },
  {
   "id": "blei2017-variational-inference",
   "title": "Variational Inference: A Review for Statisticians",
   "discipline": "statistics",
   "year": 2017,
   "authors": "David M. Blei, Alp Kucukelbir, Jon D. McAuliffe",
   "url": "papers/blei2017-variational-inference.html",
   "key_terms": [
    "variational inference",
    "Bayesian statistics",
    "posterior",
    "probability densities",
    "optimization",
    "Markov chain Monte Carlo",
    "Kullback-Leibler divergence",
    "latent variables",
    "mean-field",
    "approximate inference"
   ],
   "phenomenon": "variational inference (approximate posterior inference)",
   "claim_type": "method",
   "method": "theory",
   "statistics": [
    {
     "term": "Bayesian",
     "evidence": "This problem is especially important in Bayesian statistics, which frames all inference about unknown quantities as a calculation involving the posterior density.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/blei2017-variational-inference.html",
      "source_body_sha256": "8354a424b4989ac6534e4c338f4460660357ed8ecd5b2169f0db265cc15f622e",
      "evidence_sha256": "0de4afa7b95914c01c0a268bb71ed7e7d881a5173253274b0f7d730c9df5cd93",
      "grounding_status": "not-exactly-located"
     }
    },
    {
     "term": "correlation",
     "evidence": "This density is highly correlated, which defines its elongated shape.",
     "decision": "accepted",
     "role": "substantive-mention",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/blei2017-variational-inference.html",
      "source_body_sha256": "8354a424b4989ac6534e4c338f4460660357ed8ecd5b2169f0db265cc15f622e",
      "evidence_sha256": "a4ee47ee43573c4217d7205675b437c630b1f7651ef6d23648c0ca5ecf81ef0d",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "mean",
     "evidence": "They find that, under standard regularity conditions, the mean-field variational posterior mean of the parameters are consistent in the frequentist sense.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/blei2017-variational-inference.html",
      "source_body_sha256": "8354a424b4989ac6534e4c338f4460660357ed8ecd5b2169f0db265cc15f622e",
      "evidence_sha256": "5428ecb539b4195a135fc0f3651c9b3fbb147f1853dc02ef0b259ff6a5f28e35",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "regression",
     "evidence": "As examples we show CAVI for a mixture of Gaussians in Section 3 and for a nonconjugate linear regression in Appendix A.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/blei2017-variational-inference.html",
      "source_body_sha256": "8354a424b4989ac6534e4c338f4460660357ed8ecd5b2169f0db265cc15f622e",
      "evidence_sha256": "8a09bc6be5f9bb21189026ab4c7d8e1167c0e78ee5bd43ca8d84aaba7f9546cf",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "standard-error",
     "evidence": "They prove consistency of these estimates at the parametric rate and show asymptotic normality with asymptotically valid standard errors.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit standard error or SEM",
     "provenance": {
      "source_path": "papers/blei2017-variational-inference.html",
      "source_body_sha256": "8354a424b4989ac6534e4c338f4460660357ed8ecd5b2169f0db265cc15f622e",
      "evidence_sha256": "01500ece2d5471a9551bc3d73b99d4c5dca943df7e87075012348129a0433fca",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "variance",
     "evidence": "We do know that variational inference generally underestimates the variance of the posterior density; this is a consequence of its objective function.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/blei2017-variational-inference.html",
      "source_body_sha256": "8354a424b4989ac6534e4c338f4460660357ed8ecd5b2169f0db265cc15f622e",
      "evidence_sha256": "cb5d909cb4809818dac63ce9f888b2f985683d0f1d1ed4de02886a73beddadd1",
      "grounding_status": "not-exactly-located"
     }
    }
   ],
   "license_spdx": "arXiv-nonexclusive-distrib-1.0",
   "license_url": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
   "source_url": "https://arxiv.org/abs/1601.00670",
   "doi": "10.1080/01621459.2017.1285773",
   "full_text_redistributable": false
  },
  {
   "id": "dwork2012-fairness",
   "title": "Fairness Through Awareness",
   "discipline": "law",
   "year": 2012,
   "authors": "Cynthia Dwork, Moritz Hardt, Toniann Pitassi, Omer Reingold, Richard Zemel",
   "url": "papers/dwork2012-fairness.html",
   "key_terms": [
    "fairness",
    "classification",
    "individuals",
    "similar",
    "metric",
    "utility",
    "discrimination",
    "parity"
   ],
   "phenomenon": "algorithmic fairness (individual fairness; similar individuals treated similarly)",
   "claim_type": "method",
   "method": "mathematical-proof",
   "statistics": [
    {
     "term": "correlation",
     "evidence": "Clicking on such an ad may be strongly correlated with membership in S (even if exposure to the ad obeys statistical parity).",
     "decision": "accepted",
     "role": "substantive-mention",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/dwork2012-fairness.html",
      "source_body_sha256": "f9349bc7bbb359fb4ef458698c43333449dca2f435a1183138b62eba092f5ca8",
      "evidence_sha256": "6264a0c82afadcf36271f26ba35666d5292a4faf08c71aaf73756a4df48c5431",
      "grounding_status": "not-exactly-located"
     }
    },
    {
     "term": "statistical-significance",
     "evidence": "Moreover, the mechanism is significantly more efficient than the general linear program.",
     "decision": "rejected",
     "role": "lexical-significance",
     "reason": "significant/significantly is used without inferential-statistical meaning",
     "provenance": {
      "source_path": "papers/dwork2012-fairness.html",
      "source_body_sha256": "f9349bc7bbb359fb4ef458698c43333449dca2f435a1183138b62eba092f5ca8",
      "evidence_sha256": "78b8838b5cd7f2361da0cf9887704de7841d30edae3f1b039b9e7b535e4d7fac",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "arXiv-nonexclusive-distrib-1.0",
   "license_url": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
   "source_url": "https://arxiv.org/abs/1104.3913",
   "doi": "10.1145/2090236.2090255",
   "full_text_redistributable": false
  },
  {
   "id": "schutt2017-schnet",
   "title": "SchNet: A Continuous-Filter Convolutional Neural Network for Modeling Quantum Interactions",
   "discipline": "chemistry",
   "year": 2017,
   "authors": "Kristof T. Schütt, Pieter-Jan Kindermans, Huziel E. Sauceda, Stefan Chmiela, Alexandre Tkatchenko, Klaus-Robert Müller",
   "url": "papers/schutt2017-schnet.html",
   "key_terms": [
    "molecules",
    "atoms",
    "energy",
    "quantum",
    "convolutional",
    "forces",
    "chemical",
    "interatomic",
    "continuous-filter"
   ],
   "phenomenon": "molecular property prediction with deep learning (quantum-chemical property learning)",
   "claim_type": "method",
   "method": "computational-model",
   "statistics": [
    {
     "term": "Bayesian",
     "evidence": "Bayesian interpolation of unequally spaced time series.",
     "decision": "rejected",
     "role": "bibliography",
     "reason": "occurrence is confined to a citation/reference rather than the article body",
     "provenance": {
      "source_path": "papers/schutt2017-schnet.html",
      "source_body_sha256": "60ba3676e47e42ca237d19cf82d6535b499de1a4c4e25dd6ecadce2430a2c77d",
      "evidence_sha256": "d4c1ce1ebbe95e590d99950f97a665a72d2f04ada85686bb7fcbc7f0ac6814dc",
      "grounding_status": "verified-reference"
     }
    },
    {
     "term": "correlation",
     "evidence": "Pearson, and A.",
     "decision": "accepted",
     "role": "substantive-mention",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/schutt2017-schnet.html",
      "source_body_sha256": "60ba3676e47e42ca237d19cf82d6535b499de1a4c4e25dd6ecadce2430a2c77d",
      "evidence_sha256": "dabf3737e534f6926ec962b7899ba9be5210dddaa11d14d928363dcb2da56b85",
      "grounding_status": "verified-reference"
     }
    }
   ],
   "license_spdx": "arXiv-nonexclusive-distrib-1.0",
   "license_url": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
   "source_url": "https://arxiv.org/abs/1706.08566",
   "doi": "",
   "full_text_redistributable": false
  },
  {
   "id": "rajpurkar2017-chexnet",
   "title": "CheXNet: Radiologist-Level Pneumonia Detection on Chest X-Rays with Deep Learning",
   "discipline": "medicine",
   "year": 2017,
   "authors": "Pranav Rajpurkar, Jeremy Irvin, Andrew Y. Ng, et al.",
   "url": "papers/rajpurkar2017-chexnet.html",
   "key_terms": [
    "CheXNet",
    "radiologist",
    "pneumonia",
    "X-ray",
    "DenseNet",
    "chest",
    "disease",
    "network",
    "images"
   ],
   "phenomenon": "automated medical image diagnosis (pneumonia detection from chest X-rays with deep learning)",
   "claim_type": "method",
   "method": "computational-model",
   "statistics": [
    {
     "term": "bootstrap",
     "evidence": "We use the bootstrap to con- struct 95% bootstrap confidence intervals (CIs), cal- culating the average F1 score for both the radiologists and CheXNet on 10,000 bootstrap samples, ",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/rajpurkar2017-chexnet.html",
      "source_body_sha256": "87a2070c3be54c7b800e071d95b97868fa9850d1561015866cd2d304fd01c34e",
      "evidence_sha256": "503484d8f63f53c3b5da5244583ce6184bf7e9ea8e43b9c93d6ace6d9b0adcf2",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "confidence-interval",
     "evidence": "We use the bootstrap to con- struct 95% bootstrap confidence intervals (CIs), cal- culating the average F1 score for both the radiologists and CheXNet on 10,000 bootstrap samples, ",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/rajpurkar2017-chexnet.html",
      "source_body_sha256": "87a2070c3be54c7b800e071d95b97868fa9850d1561015866cd2d304fd01c34e",
      "evidence_sha256": "503484d8f63f53c3b5da5244583ce6184bf7e9ea8e43b9c93d6ace6d9b0adcf2",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "mean",
     "evidence": "We report the mean of the 4 resulting F1 scores for each radi- ologist and for CheXNet, along with the average F1 across the radiologists.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "numeric mean or average in main text",
     "provenance": {
      "source_path": "papers/rajpurkar2017-chexnet.html",
      "source_body_sha256": "87a2070c3be54c7b800e071d95b97868fa9850d1561015866cd2d304fd01c34e",
      "evidence_sha256": "bc05d7ea160408cfd0b483a7d9c2f654b0873ccd1a2f86fa50dfd112f1d6d0a7",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "standard-deviation",
     "evidence": "Before inputting the images into the network, we downscale the images to 224×224 and normalize based on the mean and standard deviation of images in the ImageNet training set.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/rajpurkar2017-chexnet.html",
      "source_body_sha256": "87a2070c3be54c7b800e071d95b97868fa9850d1561015866cd2d304fd01c34e",
      "evidence_sha256": "66513eccbe97d280bdd8337dc5d2048087d06d1e6b45068dca368fdd09d3aa3f",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "statistical-significance",
     "evidence": "We find that the difference in F1 scores — 0.051 (95% CI 0.005, 0.084) — does not contain 0, and therefore conclude that the performance of CheXNet is statistically significantly h",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit inferential-statistical context",
     "provenance": {
      "source_path": "papers/rajpurkar2017-chexnet.html",
      "source_body_sha256": "87a2070c3be54c7b800e071d95b97868fa9850d1561015866cd2d304fd01c34e",
      "evidence_sha256": "64c3807b731372b14026d8d1d67017d27f54ba4d6696a33f2d5f297a9136d14a",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "arXiv-nonexclusive-distrib-1.0",
   "license_url": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
   "source_url": "https://arxiv.org/abs/1711.05225",
   "doi": "",
   "full_text_redistributable": false
  },
  {
   "id": "athey2016-causal-trees",
   "title": "Recursive Partitioning for Heterogeneous Causal Effects",
   "discipline": "economics",
   "year": 2016,
   "authors": "Susan Athey, Guido Imbens",
   "url": "papers/athey2016-causal-trees.html",
   "key_terms": [
    "treatment",
    "causal",
    "effects",
    "partition",
    "honest",
    "tree",
    "covariates",
    "estimation",
    "subgroup"
   ],
   "phenomenon": "heterogeneous causal (treatment) effect estimation",
   "claim_type": "method",
   "method": "computational-model",
   "statistics": [
    {
     "term": "Bayesian",
     "evidence": "Using Bayesian nonparametric methods with Dirichlet priors, they project their estimates of heterogeneous treatment effects",
     "decision": "accepted",
     "role": "substantive-mention",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/athey2016-causal-trees.html",
      "source_body_sha256": "2052c25f66eb6a30f4481d0914caec8b81e1aa0d0e535cdb345f6a49775100e0",
      "evidence_sha256": "456e24dae42b315ac6d2705424e138f490175a20a4478844c9e9f308f3c0ae38",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "confidence-interval",
     "evidence": "The approach enables the construction of valid confidence intervals for treatment effects, even in samples with many covariates relative to the sample size, and without “sparsity” ",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/athey2016-causal-trees.html",
      "source_body_sha256": "2052c25f66eb6a30f4481d0914caec8b81e1aa0d0e535cdb345f6a49775100e0",
      "evidence_sha256": "9293b76da310eb3c98ff9c45432b0adb1b5d2672fee7d01263682cfeb8201aef",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "false-discovery-rate",
     "evidence": "For experiments, our method allows researchers to identify heterogeneity in treatment effects that was not specified in a pre-analysis plan, without concern about invalidating infe",
     "decision": "rejected",
     "role": "non-statistical-sense",
     "reason": "lexical match does not denote the controlled statistical concept",
     "provenance": {
      "source_path": "papers/athey2016-causal-trees.html",
      "source_body_sha256": "2052c25f66eb6a30f4481d0914caec8b81e1aa0d0e535cdb345f6a49775100e0",
      "evidence_sha256": "d67a0389b30cb89d413ad7943200f67f75f261d107db3f30b05ba5c87a9bc6a3",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "hypothesis-testing",
     "evidence": "or estimating heterogeneity in causal effects in exper- imental and observational studies, and for conducting hypothesis tests about the magnitude of the differences in treatment e",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit hypothesis-testing language",
     "provenance": {
      "source_path": "papers/athey2016-causal-trees.html",
      "source_body_sha256": "2052c25f66eb6a30f4481d0914caec8b81e1aa0d0e535cdb345f6a49775100e0",
      "evidence_sha256": "baf0cf82f05fff4bac9cf1b57eb78cfc524e7cef642a20198dd105b5011e046f",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "mean",
     "evidence": "The treatment effect estimator within a leaf is the same as the adaptive method, that is, the sample mean of Y ∗ i within the leaf.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "numeric mean or average in main text",
     "provenance": {
      "source_path": "papers/athey2016-causal-trees.html",
      "source_body_sha256": "2052c25f66eb6a30f4481d0914caec8b81e1aa0d0e535cdb345f6a49775100e0",
      "evidence_sha256": "3ff62790b5740a6ed009718abf6ebbbc6bfcf78ca58ca7adc079f882e187134e",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "null-hypothesis",
     "evidence": "he last estimator we look for splits with the largest value for the square of the t-statistic for testing the null hypothesis that the average treatment effect is the same in the t",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit null-hypothesis language",
     "provenance": {
      "source_path": "papers/athey2016-causal-trees.html",
      "source_body_sha256": "2052c25f66eb6a30f4481d0914caec8b81e1aa0d0e535cdb345f6a49775100e0",
      "evidence_sha256": "c1ca825f83065897d13c7effa60ec20c15dc5360c2518ef93017fae042ae28fb",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "p-value",
     "evidence": "It enables researchers to let the data discover relevant subgroups without falling prey to concerns of multiple hypothesis testing that would invalidate p-values.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit p-value or inferential p notation",
     "provenance": {
      "source_path": "papers/athey2016-causal-trees.html",
      "source_body_sha256": "2052c25f66eb6a30f4481d0914caec8b81e1aa0d0e535cdb345f6a49775100e0",
      "evidence_sha256": "2f7bfc65ccd2820514a51db18dfbafea569de8d77e75c98afee717bc01e345f5",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "regression",
     "evidence": "regression trees ([4]), random forests ([3]), LASSO ([24]), sup- port vector machines ([26]), etc.) entail building a model of the relationship between attributes and outcomes, wit",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/athey2016-causal-trees.html",
      "source_body_sha256": "2052c25f66eb6a30f4481d0914caec8b81e1aa0d0e535cdb345f6a49775100e0",
      "evidence_sha256": "8620bb7aae76a3aa12c2919dc750784950e9d611f614c26483bb07c5d49af4bb",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "sample-size",
     "evidence": "The approach enables the construction of valid confidence intervals for treatment effects, even in samples with many covariates relative to the sample size, and without “sparsity” ",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/athey2016-causal-trees.html",
      "source_body_sha256": "2052c25f66eb6a30f4481d0914caec8b81e1aa0d0e535cdb345f6a49775100e0",
      "evidence_sha256": "9293b76da310eb3c98ff9c45432b0adb1b5d2672fee7d01263682cfeb8201aef",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "t-test",
     "evidence": "4.4 Squared T-statistic Trees (TS) For the last estimator we look for splits with the largest value for the square of the t-statistic",
     "decision": "rejected",
     "role": "non-statistical-sense",
     "reason": "lexical match does not denote the controlled statistical concept",
     "provenance": {
      "source_path": "papers/athey2016-causal-trees.html",
      "source_body_sha256": "2052c25f66eb6a30f4481d0914caec8b81e1aa0d0e535cdb345f6a49775100e0",
      "evidence_sha256": "23ac0d99bf98ddf64c7a4c42395d9ff75818acf5dfca4891905ae755119bf38c",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "variance",
     "evidence": "es, anticipating that bias will be eliminated by honest estimation, but also accounting for the change in the variance of treatment effect estimates within each subpopulation as a ",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/athey2016-causal-trees.html",
      "source_body_sha256": "2052c25f66eb6a30f4481d0914caec8b81e1aa0d0e535cdb345f6a49775100e0",
      "evidence_sha256": "4f6c497bd80c19229d7f3edaeda5d0754092f3b2238a668665f6f01d10c4f497",
      "grounding_status": "not-exactly-located"
     }
    }
   ],
   "license_spdx": "arXiv-nonexclusive-distrib-1.0",
   "license_url": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
   "source_url": "https://arxiv.org/abs/1504.01132",
   "doi": "10.1073/pnas.1510489113",
   "full_text_redistributable": false
  },
  {
   "id": "pennington2014-glove",
   "title": "GloVe: Global Vectors for Word Representation",
   "discipline": "linguistics",
   "year": 2014,
   "authors": "Jeffrey Pennington, Richard Socher, Christopher D. Manning",
   "url": "papers/pennington2014-glove.html",
   "key_terms": [
    "word",
    "vectors",
    "co-occurrence",
    "corpus",
    "semantic",
    "matrix",
    "analogy",
    "GloVe"
   ],
   "phenomenon": "word embeddings (distributed vector representations of words)",
   "claim_type": "method",
   "method": "computational-model",
   "statistics": [
    {
     "term": "correlation",
     "evidence": "We compute Spearman’s rank correlation coefficient between this score and the human judgments.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/pennington2014-glove.html",
      "source_body_sha256": "f60d77e4c28c47ef2e27ed2d9668ee038b119b01a59ad9588f538cca90727def",
      "evidence_sha256": "3771cd0a8963465f66a7da87b7078de14de240ecbb4c8864c53ab3d74b8fa924",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "regression",
     "evidence": "The result is a new global log- bilinear regression model that combines the advantages of the two major model families in the literature: global matrix factorization and local cont",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/pennington2014-glove.html",
      "source_body_sha256": "f60d77e4c28c47ef2e27ed2d9668ee038b119b01a59ad9588f538cca90727def",
      "evidence_sha256": "b33f4b87787a96523d3355ae4b1520909fcbd1c3481835a1030d2ee77cfc052c",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "open-access",
   "license_url": "",
   "source_url": "https://nlp.stanford.edu/pubs/glove.pdf",
   "doi": "10.3115/v1/D14-1162",
   "full_text_redistributable": false
  },
  {
   "id": "onnela2007-tie-strengths",
   "title": "Structure and Tie Strengths in Mobile Communication Networks",
   "discipline": "sociology",
   "year": 2007,
   "authors": "J.-P. Onnela, J. Saramäki, J. Hyvönen, A.-L. Barabási, et al.",
   "url": "papers/onnela2007-tie-strengths.html",
   "key_terms": [
    "ties",
    "network",
    "communication",
    "mobile",
    "diffusion",
    "weak",
    "strong",
    "phone",
    "coupling"
   ],
   "phenomenon": "tie strength and network structure (weak ties hold social networks together)",
   "claim_type": "existence",
   "method": "computational-model",
   "statistics": [
    {
     "term": "bootstrap",
     "evidence": "(Npc LCCR f − cf obtain sets of estimates of for different thresholding schemes and sampling techniques using bootstrapping, in which we randomly choose half of the points to be in",
     "decision": "rejected",
     "role": "bibliography",
     "reason": "occurrence appears only in references",
     "provenance": {
      "source_path": "papers/onnela2007-tie-strengths.html",
      "source_body_sha256": "be6584cf8d748b576094b4ba24074505bf925ddf7673b166a352e41c2308d4af",
      "evidence_sha256": "8a60af7f02f965c48a0acb3a057eb5fe82662e93f52bfe9abb9c1a56b80f2c70",
      "grounding_status": "not-exactly-located"
     }
    },
    {
     "term": "correlation",
     "evidence": "As expected, these two variables are statistically dependent, giving rise to Pearson’s linear correlation coefficient of 0.70.",
     "decision": "accepted",
     "role": "substantive-mention",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/onnela2007-tie-strengths.html",
      "source_body_sha256": "be6584cf8d748b576094b4ba24074505bf925ddf7673b166a352e41c2308d4af",
      "evidence_sha256": "9c4ae55db1763a6556596e56cde8e4a85cb0070d7b2315413bc555659e30fa1f",
      "grounding_status": "verified-reference"
     }
    },
    {
     "term": "mean",
     "evidence": "We take as the mean of the bootstrap distribution, and the error bounds are taken as distribution ( (∞cf (∞cf 6).",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/onnela2007-tie-strengths.html",
      "source_body_sha256": "be6584cf8d748b576094b4ba24074505bf925ddf7673b166a352e41c2308d4af",
      "evidence_sha256": "8b8a765cb56a4715d0a84df33f16d8625d8b60a46aa76ba344f8758e6f786b34",
      "grounding_status": "verified-reference"
     }
    },
    {
     "term": "p-value",
     "evidence": "The green (◊), red ((cid:31)), and blue (ο) curves are associated with system sizes , ≈N , and they were obtained using resampling extraction with node occupation probabilities p=0",
     "decision": "rejected",
     "role": "non-statistical-sense",
     "reason": "lexical match does not denote the controlled statistical concept",
     "provenance": {
      "source_path": "papers/onnela2007-tie-strengths.html",
      "source_body_sha256": "be6584cf8d748b576094b4ba24074505bf925ddf7673b166a352e41c2308d4af",
      "evidence_sha256": "2a8ee24a0260d6d2809bc4799ccf91e3fc34aa87634c90d33b0c0a481c3f825d",
      "grounding_status": "verified-reference"
     }
    },
    {
     "term": "sample-size",
     "evidence": "We in different sample sizes, and in the limit of setting the network is likely to consider only the LCC of the resulting resample, since for become fragmented.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/onnela2007-tie-strengths.html",
      "source_body_sha256": "be6584cf8d748b576094b4ba24074505bf925ddf7673b166a352e41c2308d4af",
      "evidence_sha256": "e34cbbde7e67d1899e7770b0b60185da126c24b9bbc731d8a91828e662ab7eda",
      "grounding_status": "verified-reference"
     }
    },
    {
     "term": "standard-deviation",
     "evidence": "Repeating this 10000 times gives a distribution of the estimates of the value the standard deviations of the .",
     "decision": "rejected",
     "role": "bibliography",
     "reason": "occurrence appears only in references",
     "provenance": {
      "source_path": "papers/onnela2007-tie-strengths.html",
      "source_body_sha256": "be6584cf8d748b576094b4ba24074505bf925ddf7673b166a352e41c2308d4af",
      "evidence_sha256": "ea93774722d3d82a429f8f6ab86aa05d6a024fd280dd3258bd14de3ae8e63e3f",
      "grounding_status": "verified-reference"
     }
    },
    {
     "term": "statistical-significance",
     "evidence": "If the tail is approximated by a power law, the exponent γk = 8.4 is significantly higher than the value observed for land lines (γ = 2.1 for the in-degree distribution, see Ref.",
     "decision": "rejected",
     "role": "lexical-significance",
     "reason": "significant/significantly is used without inferential-statistical meaning",
     "provenance": {
      "source_path": "papers/onnela2007-tie-strengths.html",
      "source_body_sha256": "be6584cf8d748b576094b4ba24074505bf925ddf7673b166a352e41c2308d4af",
      "evidence_sha256": "292ab515684b806e2943d41840e56a28bf9daeb509195425afa5ca853166ff15",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "arxiv-open",
   "license_url": "",
   "source_url": "https://arxiv.org/abs/physics/0610104",
   "doi": "10.1073/pnas.0610245104",
   "full_text_redistributable": false
  },
  {
   "id": "hamilton2016-diachronic-embeddings",
   "title": "Diachronic Word Embeddings Reveal Statistical Laws of Semantic Change",
   "discipline": "digital-humanities",
   "year": 2016,
   "authors": "William L. Hamilton, Jure Leskovec, Dan Jurafsky",
   "url": "papers/hamilton2016-diachronic-embeddings.html",
   "key_terms": [
    "semantic",
    "change",
    "word",
    "embeddings",
    "frequency",
    "polysemous",
    "diachronic",
    "historical",
    "corpora"
   ],
   "phenomenon": "semantic change (statistical laws: conformity and innovation)",
   "claim_type": "existence",
   "method": "corpus-analysis",
   "statistics": [
    {
     "term": "Bayesian",
     "evidence": "ing: perhaps people are more likely to use rare words mistakenly in novel ways, a mechanism for- malizable by Bayesian models of word learning and corresponding to the biological n",
     "decision": "accepted",
     "role": "substantive-mention",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/hamilton2016-diachronic-embeddings.html",
      "source_body_sha256": "a8411926d2907e2a1f9621f422542e9d95f92226bd3fb3ff2c7795b047878e41",
      "evidence_sha256": "45d48640efe982f593616c7857cb0706634e9a4ed7b094146d64dc16621b91c3",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "correlation",
     "evidence": "Across datasets, we found Pearson correlations in the range 0.45 < r < 0.8 (all p < 0.05), confirm- ing frequent words tend to be used in a greater di- versity of contexts.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/hamilton2016-diachronic-embeddings.html",
      "source_body_sha256": "a8411926d2907e2a1f9621f422542e9d95f92226bd3fb3ff2c7795b047878e41",
      "evidence_sha256": "9f2e6c68b87f0f25807242cb62416ecb0a793d028a96979ce559c11e5a5418b3",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "hypothesis-testing",
     "evidence": "similarity time-series Measuring Pair-wise how the cosine-similarity between pairs of words changes over time allows us to test hypotheses about specific linguistic or cultural shi",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/hamilton2016-diachronic-embeddings.html",
      "source_body_sha256": "a8411926d2907e2a1f9621f422542e9d95f92226bd3fb3ff2c7795b047878e41",
      "evidence_sha256": "870c80e7a8e9021ee22fd2a6e44721d9357f04a69d935e01449cd698ba920798",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "p-value",
     "evidence": "e), but there were some differences in whether the methods deemed the shifts statistically significant at the p < 0.05 level.7 Overall, SGNS performed the best on the full English ",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit p-value or inferential p notation",
     "provenance": {
      "source_path": "papers/hamilton2016-diachronic-embeddings.html",
      "source_body_sha256": "a8411926d2907e2a1f9621f422542e9d95f92226bd3fb3ff2c7795b047878e41",
      "evidence_sha256": "939f1f85aef9213002ac9cbd9df5d7c4bfbb24522b21b8c06bc733f7db400233",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "regression",
     "evidence": "The plots show robust linear regression fits (Huber, 2011) with 95% CIs on the 2000s decade of the COHA lemma data.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/hamilton2016-diachronic-embeddings.html",
      "source_body_sha256": "a8411926d2907e2a1f9621f422542e9d95f92226bd3fb3ff2c7795b047878e41",
      "evidence_sha256": "9e9a80585dcda3bb7e3886173a7da38755c50d98b3eebb563988a916cdb49cff",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "sample-size",
     "evidence": "This was necessary because there was substantial changes in the median con- textual diversity scores over years due to changes in corpus sample sizes etc.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/hamilton2016-diachronic-embeddings.html",
      "source_body_sha256": "a8411926d2907e2a1f9621f422542e9d95f92226bd3fb3ff2c7795b047878e41",
      "evidence_sha256": "b7a955d8687b6caf204c94cb4662c555e0d8b941bfb9a56667a94a5883ab8dd4",
      "grounding_status": "verified-reference"
     }
    },
    {
     "term": "statistical-significance",
     "evidence": "eir Spearman correlation with time), but there were some differences in whether the methods deemed the shifts statistically significant at the p < 0.05 level.7 Overall, SGNS perfor",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit inferential-statistical context",
     "provenance": {
      "source_path": "papers/hamilton2016-diachronic-embeddings.html",
      "source_body_sha256": "a8411926d2907e2a1f9621f422542e9d95f92226bd3fb3ff2c7795b047878e41",
      "evidence_sha256": "3bed193af264c03d4618da2999dd6a8cfd7d5b234f4263a850d4cb1b4aadc712",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "variance",
     "evidence": "d- dings the effect of frequency is confounded by the fact that high frequency words have less finite- sample variance in their co-occurrence estimates, which makes the word vector",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/hamilton2016-diachronic-embeddings.html",
      "source_body_sha256": "a8411926d2907e2a1f9621f422542e9d95f92226bd3fb3ff2c7795b047878e41",
      "evidence_sha256": "072cc6476270c880747fd93e9e759cd2267c0777811a5b48e55307d808f66fc9",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "arXiv-nonexclusive-distrib-1.0",
   "license_url": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
   "source_url": "https://arxiv.org/abs/1605.09096",
   "doi": "10.18653/v1/P16-1141",
   "full_text_redistributable": false
  },
  {
   "id": "grover1996-search",
   "title": "A Fast Quantum Mechanical Algorithm for Database Search",
   "discipline": "quantum-computing",
   "year": 1996,
   "authors": "Lov K. Grover",
   "url": "papers/grover1996-search.html",
   "key_terms": [
    "quantum",
    "amplitude",
    "superposition",
    "unitary",
    "algorithm",
    "database",
    "probability",
    "states",
    "phase"
   ],
   "phenomenon": "quantum search of an unstructured database (quadratic speedup)",
   "claim_type": "method",
   "method": "mathematical-proof",
   "statistics": [
    {
     "term": "mean",
     "evidence": "It is easily seen N ---- 2 that this algorithm will need to look at an average of items before finding the desired item.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "numeric mean or average in main text",
     "provenance": {
      "source_path": "papers/grover1996-search.html",
      "source_body_sha256": "8ffb3016aced129fe62110df081d0308ea008e8445e210eca5e6768f50f8ad75",
      "evidence_sha256": "cc6b4649545c596ebbb74caba884fb56f3c2ac7ea2b6fed9e08522791d59deb5",
      "grounding_status": "not-exactly-located"
     }
    }
   ],
   "license_spdx": "arxiv-open",
   "license_url": "",
   "source_url": "https://arxiv.org/abs/quant-ph/9605043",
   "doi": "10.1145/237814.237866",
   "full_text_redistributable": false
  },
  {
   "id": "gatys2015-style-transfer",
   "title": "A Neural Algorithm of Artistic Style",
   "discipline": "art-history",
   "year": 2015,
   "authors": "Leon A. Gatys, Alexander S. Ecker, Matthias Bethge",
   "url": "papers/gatys2015-style-transfer.html",
   "key_terms": [
    "style",
    "content",
    "image",
    "convolutional",
    "feature",
    "neural",
    "representations",
    "texture",
    "Gram"
   ],
   "phenomenon": "neural style transfer (separating and recombining image content and style)",
   "claim_type": "method",
   "method": "computational-model",
   "statistics": [
    {
     "term": "standard-error",
     "evidence": "(1) (2) from which the gradient with respect to the image (cid:126)x can be computed using standard error back-propagation.",
     "decision": "rejected",
     "role": "non-statistical-sense",
     "reason": "lexical match does not denote the controlled statistical concept",
     "provenance": {
      "source_path": "papers/gatys2015-style-transfer.html",
      "source_body_sha256": "e3cf255ae829008deddd4782bbb3196d39cdc2d0c97af64bdfa3d3f96e345cb6",
      "evidence_sha256": "0a8de083a8317e464716c8618949c169ef789af9749fb851114fdf677b403571",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "arXiv-nonexclusive-distrib-1.0",
   "license_url": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
   "source_url": "https://arxiv.org/abs/1508.06576",
   "doi": "",
   "full_text_redistributable": false
  },
  {
   "id": "levine2016-visuomotor",
   "title": "End-to-End Training of Deep Visuomotor Policies",
   "discipline": "robotics",
   "year": 2016,
   "authors": "Sergey Levine, Chelsea Finn, Trevor Darrell, Pieter Abbeel",
   "url": "papers/levine2016-visuomotor.html",
   "key_terms": [
    "visuomotor policies",
    "policy search",
    "guided policy search",
    "reinforcement learning",
    "convolutional neural networks",
    "end-to-end",
    "sensorimotor",
    "torques",
    "perception",
    "supervised learning"
   ],
   "phenomenon": "robot visuomotor control (end-to-end learning of policies from pixels to torques)",
   "claim_type": "method",
   "method": "computational-model",
   "statistics": [
    {
     "term": "correlation",
     "evidence": "However, we can observe that the dynamics at nearby time steps are strongly correlated, and we can dramatically reduce the sample complexity of the dynamics fitting by bringing in ",
     "decision": "accepted",
     "role": "substantive-mention",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/levine2016-visuomotor.html",
      "source_body_sha256": "a1ac1028f7cbc6f3b460a5322eb114e4f429129893d221e23829869e4805ab9c",
      "evidence_sha256": "9a2efc9606445e27b32396089892043c4a6e52da5e524c6260eafc51dd53614b",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "mean",
     "evidence": "rgence term DKL(p(ut|xt)(cid:107)πθ(ut|xt)) inside Lp(p, θ) for iLQG, we also estimate a linearization of the mean of the conditionally Gaussian policy πθ(ut|ot) with respect to th",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/levine2016-visuomotor.html",
      "source_body_sha256": "a1ac1028f7cbc6f3b460a5322eb114e4f429129893d221e23829869e4805ab9c",
      "evidence_sha256": "cedec4f6f7adea8cc378d789c7af4bdf9351fa6de15740212455cb724b94a62e",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "regression",
     "evidence": "A simple way to fit these linear-Gaussian dynamics is to use linear regression to determine fx, fu, and fc, and fit Ft based on the errors.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/levine2016-visuomotor.html",
      "source_body_sha256": "a1ac1028f7cbc6f3b460a5322eb114e4f429129893d221e23829869e4805ab9c",
      "evidence_sha256": "b02b7501f05efe648336b9fbe88e4a76109cfd7a4525c0464afb5b02f924a72e",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "standard-deviation",
     "evidence": "Error bars indicate one standard deviation.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/levine2016-visuomotor.html",
      "source_body_sha256": "a1ac1028f7cbc6f3b460a5322eb114e4f429129893d221e23829869e4805ab9c",
      "evidence_sha256": "e7fa4ebb099911ca45f277f296fa3b7e25dc26de8f31b23775890b0dff842106",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "arXiv-nonexclusive-distrib-1.0",
   "license_url": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
   "source_url": "https://arxiv.org/abs/1504.00702",
   "doi": "",
   "full_text_redistributable": false
  },
  {
   "id": "novoselov2005-dirac-fermions",
   "title": "Two-Dimensional Gas of Massless Dirac Fermions in Graphene",
   "discipline": "materials-science",
   "year": 2005,
   "authors": "K. S. Novoselov, A. K. Geim, S. V. Morozov, et al.",
   "url": "papers/novoselov2005-dirac-fermions.html",
   "key_terms": [
    "graphene",
    "Dirac",
    "massless",
    "carriers",
    "quantum Hall",
    "relativistic",
    "mobility",
    "fermions",
    "half-integer"
   ],
   "phenomenon": "massless Dirac fermions in graphene (relativistic quasiparticles in a solid)",
   "claim_type": "existence",
   "method": "experiment",
   "statistics": [
    {
     "term": "mean",
     "evidence": "The histogram (orange background) shows the number P of devices exhibiting ρmax within 10% intervals around the average value of ≈h/4e2.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "numeric mean or average in main text",
     "provenance": {
      "source_path": "papers/novoselov2005-dirac-fermions.html",
      "source_body_sha256": "125f897052288d48f33b05799fae5a8d8d5918f50b5d8f7c4040dc2c133dbbc8",
      "evidence_sha256": "f4944495933fb45aa8b7234653ed12e30be89e7785fb56c48ee175bb5486da85",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "arxiv-open",
   "license_url": "",
   "source_url": "https://arxiv.org/abs/cond-mat/0509330",
   "doi": "10.1038/nature04233",
   "full_text_redistributable": false
  },
  {
   "id": "love2014-deseq2",
   "title": "Moderated Estimation of Fold Change and Dispersion for RNA-seq Data with DESeq2",
   "discipline": "genetics",
   "year": 2014,
   "authors": "Michael I. Love, Wolfgang Huber, Simon Anders",
   "url": "papers/love2014-deseq2.html",
   "key_terms": [
    "RNA-seq",
    "differential expression",
    "dispersion",
    "fold change",
    "shrinkage estimation",
    "empirical Bayes",
    "negative binomial",
    "count data",
    "generalized linear model",
    "maximum a posteriori",
    "read counts"
   ],
   "phenomenon": "differential expression analysis of RNA-seq count data",
   "claim_type": "method",
   "method": "computational-model",
   "statistics": [
    {
     "term": "Bayesian",
     "evidence": "DSS [6] uses a Bayesian approach to provide an estimate for the dispersion for individual genes that accounts for the heterogeneity of dispersion values for different genes.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/love2014-deseq2.html",
      "source_body_sha256": "c3e9435c29c373f44bf5737dce5832e1ba90c0af99e0f330ad52253006268c45",
      "evidence_sha256": "6225232a07b790138d7cd7b083ad1236af8f2b6e7d21c45b96653fa38c283816",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "R-squared",
     "evidence": "The adjusted Rand index [37] was used to compare a hierarchical clustering based on various distances with the true cluster membership.",
     "decision": "rejected",
     "role": "non-statistical-sense",
     "reason": "lexical match does not denote the controlled statistical concept",
     "provenance": {
      "source_path": "papers/love2014-deseq2.html",
      "source_body_sha256": "c3e9435c29c373f44bf5737dce5832e1ba90c0af99e0f330ad52253006268c45",
      "evidence_sha256": "2becc104302e3ad48b490412158f56c1479f5f0e652c6471d36f2195c05b619c",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "chi-square",
     "evidence": "The sampling distribution of a dispersion estimator is approx- imately a scaled χ 2 distribution with m − p degrees of freedom, with m the number of samples and p the number of coe",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/love2014-deseq2.html",
      "source_body_sha256": "c3e9435c29c373f44bf5737dce5832e1ba90c0af99e0f330ad52253006268c45",
      "evidence_sha256": "ef07a262e9e4d880b77f17304f322f130703f038c832a0f9a07da3a4655656af",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "correlation",
     "evidence": "If the size factors are not equal across samples, but not correlated with condition.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/love2014-deseq2.html",
      "source_body_sha256": "c3e9435c29c373f44bf5737dce5832e1ba90c0af99e0f330ad52253006268c45",
      "evidence_sha256": "fa4e9ca0203e718a52b32b3481a7e8b5e9ee4626a8961216453b49d192793750",
      "grounding_status": "not-exactly-located"
     }
    },
    {
     "term": "degrees-of-freedom",
     "evidence": "The sampling distribution of a dispersion estimator is approx- imately a scaled χ 2 distribution with m − p degrees of freedom, with m the number of samples and p the number of coe",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "degrees of freedom in a statistical model/test",
     "provenance": {
      "source_path": "papers/love2014-deseq2.html",
      "source_body_sha256": "c3e9435c29c373f44bf5737dce5832e1ba90c0af99e0f330ad52253006268c45",
      "evidence_sha256": "ef07a262e9e4d880b77f17304f322f130703f038c832a0f9a07da3a4655656af",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "effect-size",
     "evidence": "eveloped a statistical framework to facilitate gene ranking and visu- alization based on stable estimation of effect sizes (LFCs), as well as testing of differential expression wit",
     "decision": "accepted",
     "role": "substantive-mention",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/love2014-deseq2.html",
      "source_body_sha256": "c3e9435c29c373f44bf5737dce5832e1ba90c0af99e0f330ad52253006268c45",
      "evidence_sha256": "ddc7e095dbc38b51f2b6efc317d89a67b5adf6d3ef641d8e7c56ca56373bb1a1",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "false-discovery-rate",
     "evidence": "Abbreviations FDR: False discovery rate; GLM: Generalized linear model; HTS: High-throughput sequencing; LFC: Logarithmic fold change; MAP: Maximum a",
     "decision": "accepted",
     "role": "substantive-mention",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/love2014-deseq2.html",
      "source_body_sha256": "c3e9435c29c373f44bf5737dce5832e1ba90c0af99e0f330ad52253006268c45",
      "evidence_sha256": "eb50f63f1d59c1de1667c67fbe7fbc8835910e7d1e25c081a468c68fae1d0b75",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "hypothesis-testing",
     "evidence": "runken fold changes and their estimates of standard error, including improved gene ranking and visualization, hypothesis tests above and below a threshold, and the regularized loga",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit hypothesis-testing language",
     "provenance": {
      "source_path": "papers/love2014-deseq2.html",
      "source_body_sha256": "c3e9435c29c373f44bf5737dce5832e1ba90c0af99e0f330ad52253006268c45",
      "evidence_sha256": "2149e2ed216ccda63624326d9c438894f3a70378957c017d2fb457d7e7b31fc0",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "mean",
     "evidence": "1 and the rlog transformation, show- ing that the rlog both stabilizes the variance through the range of the mean of counts and helps to find meaningful patterns in the data.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/love2014-deseq2.html",
      "source_body_sha256": "c3e9435c29c373f44bf5737dce5832e1ba90c0af99e0f330ad52253006268c45",
      "evidence_sha256": "539d9b8af90c4e5e13af1486075def819e050632f9cd4943dd91699f95ee17ed",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "median",
     "evidence": "The median sensitivity estimates were typically between 0.2 and 0.4 for all algorithms.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/love2014-deseq2.html",
      "source_body_sha256": "c3e9435c29c373f44bf5737dce5832e1ba90c0af99e0f330ad52253006268c45",
      "evidence_sha256": "391fa87c83ac4efdf53731401f68b01ea1c69cfc5cbb9a80a98f7e8e39731724",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "null-hypothesis",
     "evidence": "The most common approach in the comparative anal- ysis of transcriptomics data is to test the null hypothesis that the logarithmic fold change (LFC) between treat- ment and control",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit null-hypothesis language",
     "provenance": {
      "source_path": "papers/love2014-deseq2.html",
      "source_body_sha256": "c3e9435c29c373f44bf5737dce5832e1ba90c0af99e0f330ad52253006268c45",
      "evidence_sha256": "f9f8c1b834d6dab8641994b4aa9bd58955d98c5255b68988751608ac4d0a28e3",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "p-value",
     "evidence": "Often the goal of differential analysis is to produce a list of genes passing multiple-test adjustment, ranked by P value.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit p-value or inferential p notation",
     "provenance": {
      "source_path": "papers/love2014-deseq2.html",
      "source_body_sha256": "c3e9435c29c373f44bf5737dce5832e1ba90c0af99e0f330ad52253006268c45",
      "evidence_sha256": "b903a0d317f9cf80cf4c7dd41f12b5cb521a14c3c68b07f4ccf45a69accdd0d3",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "regression",
     "evidence": "We do not use ordinary least-squares regression but rather gamma-family GLM regression.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/love2014-deseq2.html",
      "source_body_sha256": "c3e9435c29c373f44bf5737dce5832e1ba90c0af99e0f330ad52253006268c45",
      "evidence_sha256": "df81eaf76c692e3a1c3a40deb75cc81d221b56bdd9d6829c38b87f0057c22f7c",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "sample-size",
     "evidence": "Furthermore, the number of genes called significantly differentially expressed depends as much on the sample size and other aspects of experimen- tal design as it does on the biolo",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/love2014-deseq2.html",
      "source_body_sha256": "c3e9435c29c373f44bf5737dce5832e1ba90c0af99e0f330ad52253006268c45",
      "evidence_sha256": "3188555b72074eabd7150153d9af09210aec91eb1614d429d7f4c984a617e6f9",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "standard-deviation",
     "evidence": "DESeq2 handles these cases by using the gene-wise estimate instead of the shrunken estimate when the former is more than 2 residual standard deviations above the curve.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/love2014-deseq2.html",
      "source_body_sha256": "c3e9435c29c373f44bf5737dce5832e1ba90c0af99e0f330ad52253006268c45",
      "evidence_sha256": "8399f845d2bb48296b78ed54548c27daef63cbe96abd91d9fa218364e51466fc",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "standard-error",
     "evidence": "features by describ- ing a number of applications possible with shrunken fold changes and their estimates of standard error, including improved gene ranking and visualization, hypo",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit standard error or SEM",
     "provenance": {
      "source_path": "papers/love2014-deseq2.html",
      "source_body_sha256": "c3e9435c29c373f44bf5737dce5832e1ba90c0af99e0f330ad52253006268c45",
      "evidence_sha256": "e03c55f56dd7fe32f327a6207d870086b040c170b7f61432925eb8ed6a068402",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "statistical-power",
     "evidence": "We furthermore compare DESeq2’s statistical power with existing tools, revealing that our methodology has high sensitivity and precision, while controlling the false positive rate.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/love2014-deseq2.html",
      "source_body_sha256": "c3e9435c29c373f44bf5737dce5832e1ba90c0af99e0f330ad52253006268c45",
      "evidence_sha256": "dcb81f67a6a6bf0c9bb39c9ede93f4aec3b654dc868d9324b6d8d335feef2d89",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "statistical-significance",
     "evidence": "For small-scale experiments, statistical significance is often a much stricter require- ment than biological significance, thereby relieving the researcher from the need to decide ",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit inferential-statistical context",
     "provenance": {
      "source_path": "papers/love2014-deseq2.html",
      "source_body_sha256": "c3e9435c29c373f44bf5737dce5832e1ba90c0af99e0f330ad52253006268c45",
      "evidence_sha256": "8e275403f4413966f1c92a9e244adfc86906ab7681e3f3fb04c8aca4a1d88586",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "variance",
     "evidence": "Empirical Bayes shrinkage for dispersion estimation Within-group variability, i.e., the variability between repli- cates, is modeled by the dispersion parameter αi, which describes",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/love2014-deseq2.html",
      "source_body_sha256": "c3e9435c29c373f44bf5737dce5832e1ba90c0af99e0f330ad52253006268c45",
      "evidence_sha256": "dc0ccb00ecfdf4defa0b1dfdf0418ff4b8d2f819fce1e7039cc2054dd4ee0f04",
      "grounding_status": "not-exactly-located"
     }
    }
   ],
   "license_spdx": "CC-BY-4.0",
   "license_url": "http://creativecommons.org/licenses/by/4.0/",
   "source_url": "https://genomebiology.biomedcentral.com/articles/10.1186/s13059-014-0550-8",
   "doi": "10.1186/s13059-014-0550-8",
   "full_text_redistributable": true
  },
  {
   "id": "mnih2013-dqn",
   "title": "Playing Atari with Deep Reinforcement Learning",
   "discipline": "artificial-intelligence",
   "year": 2013,
   "authors": "Volodymyr Mnih, Koray Kavukcuoglu, David Silver, et al.",
   "url": "papers/mnih2013-dqn.html",
   "key_terms": [
    "reinforcement",
    "learning",
    "Q-learning",
    "pixels",
    "reward",
    "value",
    "games",
    "convolutional",
    "policy"
   ],
   "phenomenon": "reinforcement learning from high-dimensional sensory input (control policies from pixels)",
   "claim_type": "method",
   "method": "computational-model",
   "statistics": [
    {
     "term": "Bayesian",
     "evidence": "Bayesian learning of recursively fac- In Proceedings of the Thirtieth International Conference on Machine tored environments.",
     "decision": "rejected",
     "role": "bibliography",
     "reason": "occurrence is confined to a citation/reference rather than the article body",
     "provenance": {
      "source_path": "papers/mnih2013-dqn.html",
      "source_body_sha256": "e2781e71fd318a90a5a5227b5a9b84fe9c78dc4d11c20b2d74ef89c25dbdb59c",
      "evidence_sha256": "bbb15e352525aa8657a7a0fa772172640aca08f0ee407ba8b20e9b5685724461",
      "grounding_status": "verified-reference"
     }
    },
    {
     "term": "correlation",
     "evidence": "Another issue is that most deep learning algorithms assume the data samples to be independent, while in reinforcement learning one typically encounters sequences of highly correlat",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/mnih2013-dqn.html",
      "source_body_sha256": "e2781e71fd318a90a5a5227b5a9b84fe9c78dc4d11c20b2d74ef89c25dbdb59c",
      "evidence_sha256": "e01f9f6e82fb552c2d42eedd9b8337f4694e187d8b062d316e3cf3ebaaa625cd",
      "grounding_status": "not-exactly-located"
     }
    },
    {
     "term": "mean",
     "evidence": "We collect a fixed set of states by running a random policy before training starts and track the average of the maximum2 predicted Q for these states.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "numeric mean or average in main text",
     "provenance": {
      "source_path": "papers/mnih2013-dqn.html",
      "source_body_sha256": "e2781e71fd318a90a5a5227b5a9b84fe9c78dc4d11c20b2d74ef89c25dbdb59c",
      "evidence_sha256": "a3d217811d4463b8d6507ffcc7b4b6c388137b0c02abfc39c9cb903ac7e44d20",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "variance",
     "evidence": "ecutive samples is inefficient, due to the strong correlations between the samples; randomizing the samples breaks these correlations and therefore reduces the variance of the upda",
     "decision": "accepted",
     "role": "substantive-mention",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/mnih2013-dqn.html",
      "source_body_sha256": "e2781e71fd318a90a5a5227b5a9b84fe9c78dc4d11c20b2d74ef89c25dbdb59c",
      "evidence_sha256": "e42191c2eee30747a4f48f9d38c3bbbc5a5823db099cffe903351d6a973bd751",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "arXiv-nonexclusive-distrib-1.0",
   "license_url": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
   "source_url": "https://arxiv.org/abs/1312.5602",
   "doi": "",
   "full_text_redistributable": false
  },
  {
   "id": "chalmers1995-hard-problem",
   "title": "Facing Up to the Problem of Consciousness",
   "discipline": "philosophy",
   "year": 1995,
   "authors": "David J. Chalmers",
   "url": "papers/chalmers1995-hard-problem.html",
   "key_terms": [
    "consciousness",
    "experience",
    "hard problem",
    "easy problems",
    "qualia",
    "reductive explanation",
    "functional explanation",
    "subjective aspect",
    "cognitive science",
    "awareness"
   ],
   "phenomenon": "consciousness (the hard problem - subjective experience)",
   "claim_type": "critique",
   "method": "interpretive",
   "statistics": [],
   "license_spdx": "open-access",
   "license_url": "",
   "source_url": "https://consc.net/papers/facing.pdf",
   "doi": "",
   "full_text_redistributable": false
  },
  {
   "id": "miller1956-magical-number",
   "title": "The Magical Number Seven, Plus or Minus Two: Some Limits on Our Capacity for Processing Information",
   "discipline": "psychology",
   "year": 1956,
   "authors": "George A. Miller",
   "url": "papers/miller1956-magical-number.html",
   "key_terms": [
    "information",
    "bits",
    "judgment",
    "memory",
    "capacity",
    "seven",
    "stimuli",
    "channel",
    "chunk"
   ],
   "phenomenon": "channel capacity / span of immediate memory (the ~7 limit)",
   "claim_type": "existence",
   "method": "experiment",
   "statistics": [
    {
     "term": "mean",
     "evidence": "Eriksen ( 5 ) reports that, when size, brightness, and hue all vary together in perfect correlation, the transmitted information is 4.1 bits as compared with an average of about 2.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/miller1956-magical-number.html",
      "source_body_sha256": "d58b6d903c1bb2d79a0124e6798b32255404387bfe4427c60911f5998f11e74a",
      "evidence_sha256": "7c900710646f0cc4d18d7e75087124d0a91b0b9579ec0441d06ccf874a34e9ee",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "standard-deviation",
     "evidence": "If I take the best estimates I can get of the channel capacities for all the stimulus variables I have mentioned, the mean is 2.6 bits and the standard deviation is only 0.6 bit.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/miller1956-magical-number.html",
      "source_body_sha256": "d58b6d903c1bb2d79a0124e6798b32255404387bfe4427c60911f5998f11e74a",
      "evidence_sha256": "931b44ad816a9823d2aec0d69fd4631731410c9096b415fbb7bb75a31b6b95de",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "statistical-significance",
     "evidence": "Since these two studies were done in different laboratories with slightly different techniques and methods of analysis, we are not in a good position to argue whether five loudness",
     "decision": "rejected",
     "role": "lexical-significance",
     "reason": "significant/significantly is used without inferential-statistical meaning",
     "provenance": {
      "source_path": "papers/miller1956-magical-number.html",
      "source_body_sha256": "d58b6d903c1bb2d79a0124e6798b32255404387bfe4427c60911f5998f11e74a",
      "evidence_sha256": "760287c224aadc5e34881abc09118d1db1540f0007959c60c448840f6784c900",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "variance",
     "evidence": "The similarity of variance and amount of information might be explained this way: When we have a large variance, we are very ignorant about what is going to happen.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/miller1956-magical-number.html",
      "source_body_sha256": "d58b6d903c1bb2d79a0124e6798b32255404387bfe4427c60911f5998f11e74a",
      "evidence_sha256": "53da200847b9614e0a6faf54f40abe925d165f230e0b8c2496c06f12f5db2d6e",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "open-access",
   "license_url": "",
   "source_url": "https://psychclassics.yorku.ca/Miller/",
   "doi": "10.1037/h0043158",
   "full_text_redistributable": false
  },
  {
   "id": "tkacik2016-information-living",
   "title": "Information Processing in Living Systems",
   "discipline": "biology",
   "year": 2016,
   "authors": "Gašper Tkačik, William Bialek",
   "url": "papers/tkacik2016-information-living.html",
   "key_terms": [
    "information theory",
    "mutual information",
    "entropy",
    "channel capacity",
    "biological systems",
    "information flow",
    "optimization",
    "noise",
    "communication channel",
    "transcription factor"
   ],
   "phenomenon": "channel capacity (information limits and optimization in living systems)",
   "claim_type": "existence",
   "method": "theory",
   "statistics": [
    {
     "term": "Bayesian",
     "evidence": "This idea can be made concrete in a Bayesian approach, placing a prior on the space of prob- ability distributions; priors that are uniform in entropy, as opposed to uniform in pro",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/tkacik2016-information-living.html",
      "source_body_sha256": "5382d6f4ae00c89fc45fe9b441359932e19e53046f7e97c73beea389dbab057d",
      "evidence_sha256": "31a3992ee680e6bec440841d5ad51948628f1a2a9eac96ac300e9fcd60f258b0",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "correlation",
     "evidence": "If x and y are jointly Gaussian, then I(x; y) = 1 2 log2(1 c2), (3) where c is the usual linear correlation coefficient.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/tkacik2016-information-living.html",
      "source_body_sha256": "5382d6f4ae00c89fc45fe9b441359932e19e53046f7e97c73beea389dbab057d",
      "evidence_sha256": "f64ab886b8b2f65e9a11fa27a122cb82fdbfbe0a62f1bc4950d91b4e7d410d98",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "mean",
     "evidence": "f’ pathways, which 11 respond with more action potentials when the light level rises above or below its local average value; Ratliff et al [135] have argued that the preponderance ",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/tkacik2016-information-living.html",
      "source_body_sha256": "5382d6f4ae00c89fc45fe9b441359932e19e53046f7e97c73beea389dbab057d",
      "evidence_sha256": "4b4fb337b61cf8839c5e14071e88d2e04dc6fc0f92aedc07da759ed863ef430e",
      "grounding_status": "not-exactly-located"
     }
    }
   ],
   "license_spdx": "arXiv-nonexclusive-distrib-1.0",
   "license_url": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
   "source_url": "https://arxiv.org/abs/1412.8752",
   "doi": "10.1146/annurev-conmatphys-031214-014803",
   "full_text_redistributable": false
  },
  {
   "id": "greentao2004-primes",
   "title": "The Primes Contain Arbitrarily Long Arithmetic Progressions",
   "discipline": "mathematics",
   "year": 2004,
   "authors": "Ben Green, Terence Tao",
   "url": "papers/greentao2004-primes.html",
   "key_terms": [
    "primes",
    "progressions",
    "arithmetic",
    "density",
    "pseudorandom",
    "integers",
    "theorem",
    "measure",
    "transference"
   ],
   "phenomenon": "arithmetic progressions in the primes (Green-Tao theorem)",
   "claim_type": "existence",
   "method": "mathematical-proof",
   "statistics": [
    {
     "term": "mean",
     "evidence": "We observe that (1 + Om(p− 1/2)) 6 (1 + p− 1/2) Om(1) ∆ Yp | 16i<j6m (cid:18) Yp hi− By the arithmetic mean-geometric mean inequality (absorbing all constants into the 1/2)Om(1) fo",
     "decision": "rejected",
     "role": "non-statistical-sense",
     "reason": "lexical match does not denote the controlled statistical concept",
     "provenance": {
      "source_path": "papers/greentao2004-primes.html",
      "source_body_sha256": "ad50f6a89bce6400e6faab8c70d049c7898e429afb1bcbd48cd45280f6c26933",
      "evidence_sha256": "bd08257165310262bf9ef3b9fad72feae97f1877e2b1bda06d3aa8916a26f60d",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "arxiv-open",
   "license_url": "",
   "source_url": "https://arxiv.org/abs/math/0404188",
   "doi": "10.4007/annals.2008.167.481",
   "full_text_redistributable": false
  },
  {
   "id": "kingma2013-vae",
   "title": "Auto-Encoding Variational Bayes",
   "discipline": "artificial-intelligence",
   "year": 2014,
   "authors": "Diederik P. Kingma, Max Welling",
   "url": "papers/kingma2013-vae.html",
   "key_terms": [
    "variational",
    "latent",
    "posterior",
    "inference",
    "estimator",
    "reparameterization",
    "gradient",
    "likelihood"
   ],
   "phenomenon": "variational autoencoder (deep generative model with latent variables; variational Bayes)",
   "claim_type": "method",
   "method": "computational-model",
   "statistics": [
    {
     "term": "Bayesian",
     "evidence": "whose continuous latent variables and/or parameters have intractable posterior distributions? The variational Bayesian (VB) approach involves the optimization of an approximation t",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/kingma2013-vae.html",
      "source_body_sha256": "022c9d7502bb0965af822a5a30d4c02306b328628721adbace6e81ee9a1e9da5",
      "evidence_sha256": "08620e3545717e95813e7319ceeed7a4726dfa3f1ca1a099ff379a7211ed577d",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "chi-square",
     "evidence": "Examples: Log-Normal (exponentiation of normally distributed variable), Gamma (a sum over exponentially distributed variables), Dirichlet (weighted sum of Gamma variates), Beta, Ch",
     "decision": "accepted",
     "role": "substantive-mention",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/kingma2013-vae.html",
      "source_body_sha256": "022c9d7502bb0965af822a5a30d4c02306b328628721adbace6e81ee9a1e9da5",
      "evidence_sha256": "2b9cf71bd6249c8512555ffe32bdaf3aa66a1468ab3fd3b2d9e50348be413cdd",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "regression",
     "evidence": "Fixed-form variational posterior approximation through stochastic linear regression.",
     "decision": "rejected",
     "role": "bibliography",
     "reason": "occurrence is confined to a citation/reference rather than the article body",
     "provenance": {
      "source_path": "papers/kingma2013-vae.html",
      "source_body_sha256": "022c9d7502bb0965af822a5a30d4c02306b328628721adbace6e81ee9a1e9da5",
      "evidence_sha256": "968276c08dc30529b898b9531b7c141931a425e9b820f42bb8e2c22c65681359",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "variance",
     "evidence": "This gradient estimator exhibits exhibits very high variance (see e.g.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/kingma2013-vae.html",
      "source_body_sha256": "022c9d7502bb0965af822a5a30d4c02306b328628721adbace6e81ee9a1e9da5",
      "evidence_sha256": "8ecb0ef6d0aa369a618d0d567fc026fbf28feabb0bac0e7014d0041796b50d71",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "arXiv-nonexclusive-distrib-1.0",
   "license_url": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
   "source_url": "https://arxiv.org/abs/1312.6114",
   "doi": "",
   "full_text_redistributable": false
  },
  {
   "id": "wager2018-causal-forests",
   "title": "Estimation and Inference of Heterogeneous Treatment Effects using Random Forests",
   "discipline": "statistics",
   "year": 2018,
   "authors": "Stefan Wager, Susan Athey",
   "url": "papers/wager2018-causal-forests.html",
   "key_terms": [
    "causal forest",
    "heterogeneous treatment effects",
    "random forests",
    "potential outcomes",
    "unconfoundedness",
    "asymptotic normality",
    "confidence intervals",
    "causal inference",
    "nonparametric",
    "covariates",
    "consistency"
   ],
   "phenomenon": "heterogeneous causal (treatment) effect estimation",
   "claim_type": "method",
   "method": "computational-model",
   "statistics": [
    {
     "term": "ANOVA",
     "evidence": "l for random forests introduced by Lin and Jeon [2006]; the key technical device used in Section 3.3.2 is the ANOVA decomposition of Efron and Stein [1981].",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/wager2018-causal-forests.html",
      "source_body_sha256": "53dd514ebf032446df15a60f38967f12613c9deba5de8adb3447d16b9bc67f50",
      "evidence_sha256": "f1fc28cb4ad189b20a9bd3cbcdb7acd9c1307a505a303bb44c3fec3dcaf4e05a",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "Bayesian",
     "evidence": "These papers use the Bayesian Additive Regression Tree (BART) method of Chipman et al.",
     "decision": "accepted",
     "role": "substantive-mention",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/wager2018-causal-forests.html",
      "source_body_sha256": "53dd514ebf032446df15a60f38967f12613c9deba5de8adb3447d16b9bc67f50",
      "evidence_sha256": "77aa14c9824a0c9e04f74717b7c2837d33edd3798127df80a47157f5e388a274",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "bootstrap",
     "evidence": "Instead of using subsampling, Breiman originally described random forests in terms of bootstrap sampling, or bagging [Breiman, 1996].",
     "decision": "accepted",
     "role": "substantive-mention",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/wager2018-causal-forests.html",
      "source_body_sha256": "53dd514ebf032446df15a60f38967f12613c9deba5de8adb3447d16b9bc67f50",
      "evidence_sha256": "930f45489a8ccf86981aa896950d1ad9f26349500f09d3e2222882884190d3bd",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "confidence-interval",
     "evidence": "We also discuss a prac- tical method for constructing asymptotic confidence intervals for the true treatment effect that are centered at the causal forest estimates.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/wager2018-causal-forests.html",
      "source_body_sha256": "53dd514ebf032446df15a60f38967f12613c9deba5de8adb3447d16b9bc67f50",
      "evidence_sha256": "d883813956766899ae09fd6ff5dc081328fa4a5b9ba24837be660c539a607f7c",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "hypothesis-testing",
     "evidence": "Quantifying uncertainty in random forests via confidence intervals and hypothesis tests.",
     "decision": "rejected",
     "role": "bibliography",
     "reason": "occurrence appears only in references",
     "provenance": {
      "source_path": "papers/wager2018-causal-forests.html",
      "source_body_sha256": "53dd514ebf032446df15a60f38967f12613c9deba5de8adb3447d16b9bc67f50",
      "evidence_sha256": "013dedcc311421e8c7df4a63784665c632c2d6b383c85b6a3f0edadf1ee475ba",
      "grounding_status": "verified-reference"
     }
    },
    {
     "term": "mean",
     "evidence": "om forests are related to kernels and nearest-neighbor methods in that they make predictions using a weighted average of “nearby” observations; however, random forests differ in th",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/wager2018-causal-forests.html",
      "source_body_sha256": "53dd514ebf032446df15a60f38967f12613c9deba5de8adb3447d16b9bc67f50",
      "evidence_sha256": "9ccc18c2f5c71b52d30e1bf60ca022263335e2d1c67d3d0ead4dfaf3d9961ad6",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "regression",
     "evidence": "Propensity score estimation: neu- ral networks, support vector machines, decision trees (CART), and meta-classifiers as alternatives to logistic regression.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/wager2018-causal-forests.html",
      "source_body_sha256": "53dd514ebf032446df15a60f38967f12613c9deba5de8adb3447d16b9bc67f50",
      "evidence_sha256": "376afa37ef1709ab7c179e3b00f91fa266b3314239f0b334e71f4b82cf10eec4",
      "grounding_status": "verified-reference"
     }
    },
    {
     "term": "sample-size",
     "evidence": "nd classical analysis tools going back to Hoeffding [1948] and H ́ajek [1968], we show that—provided the sub- sample size s scales appropriately with n—the predictions made by a ca",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/wager2018-causal-forests.html",
      "source_body_sha256": "53dd514ebf032446df15a60f38967f12613c9deba5de8adb3447d16b9bc67f50",
      "evidence_sha256": "b587a8d61291bc22c41f58c8118e6fca0738d28fae92e5ff9a409a9671745cc7",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "standard-error",
     "evidence": "Standard errors for bagged and random forest estimators.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/wager2018-causal-forests.html",
      "source_body_sha256": "53dd514ebf032446df15a60f38967f12613c9deba5de8adb3447d16b9bc67f50",
      "evidence_sha256": "c2c265889b3698d4003a9be036179c93e9123b4b39d090e727572866ba5cdc13",
      "grounding_status": "not-exactly-located"
     }
    },
    {
     "term": "variance",
     "evidence": "the subsample size s scales as s (cid:16) nβ for some βmin < β < 1 (7) Moreover, we show that the asymptotic variance of causal forests can be accurately estimated.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/wager2018-causal-forests.html",
      "source_body_sha256": "53dd514ebf032446df15a60f38967f12613c9deba5de8adb3447d16b9bc67f50",
      "evidence_sha256": "a3e0f56a6b30b82a78aa1368a92c315a8135243a003d9cc675f9835a2c8e33f9",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "arXiv-nonexclusive-distrib-1.0",
   "license_url": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
   "source_url": "https://arxiv.org/abs/1510.04342",
   "doi": "10.1080/01621459.2017.1319839",
   "full_text_redistributable": false
  },
  {
   "id": "planck2018-cosmology",
   "title": "Planck 2018 Results. VI. Cosmological Parameters",
   "discipline": "astronomy",
   "year": 2020,
   "authors": "Planck Collaboration (N. Aghanim et al.)",
   "url": "papers/planck2018-cosmology.html",
   "key_terms": [
    "cosmological parameters",
    "Planck",
    "cosmic microwave background",
    "polarization",
    "Hubble constant",
    "dark matter",
    "baryon density",
    "lensing",
    "dark energy",
    "reionization"
   ],
   "phenomenon": "cosmological parameters (base-LCDM cosmology from the CMB)",
   "claim_type": "existence",
   "method": "observation",
   "statistics": [
    {
     "term": "Bayesian",
     "evidence": "Phys., 102, 48, arXiv:1804.09678 Cardona, W., Kunz, M., & Pettorino, V., Determining H0 with Bayesian hyper- parameters.",
     "decision": "rejected",
     "role": "bibliography",
     "reason": "occurrence is confined to a citation/reference rather than the article body",
     "provenance": {
      "source_path": "papers/planck2018-cosmology.html",
      "source_body_sha256": "3d9ded3ff4f8608bf90fa832d09d397b49eb0a1e3c0de1ff46867c52e90a2b73",
      "evidence_sha256": "abd16df322c955f97630ae53e78865b2763ae370857aa2705d469330bdebfc05",
      "grounding_status": "verified-reference"
     }
    },
    {
     "term": "chi-square",
     "evidence": "nsfer function errors, we recalibrate the T E and EE spectra against a fiducial theoretical model to minimize χ2 = (CD − GCTh)M−1(CD − GCTh), (2a) ν parameters contained in the dia",
     "decision": "accepted",
     "role": "substantive-mention",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/planck2018-cosmology.html",
      "source_body_sha256": "3d9ded3ff4f8608bf90fa832d09d397b49eb0a1e3c0de1ff46867c52e90a2b73",
      "evidence_sha256": "bdca0b4493bfb8b1176de4cd8fda677fb296e54104e018a10e254fd5aca1d1ca",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "confidence-interval",
     "evidence": "The shaded bands are mid- dle 68th and 95th percentiles (note that this does not correspond exactly to confidence intervals).",
     "decision": "accepted",
     "role": "substantive-mention",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/planck2018-cosmology.html",
      "source_body_sha256": "3d9ded3ff4f8608bf90fa832d09d397b49eb0a1e3c0de1ff46867c52e90a2b73",
      "evidence_sha256": "7815141bad6718503a9ab1a89290aa445cffbd9b1c1612ece3a266d395877233",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "correlation",
     "evidence": "We find a 0.3 % con- straint from Planck: Ωmh3 = 0.09633 ± 0.00029 (68 %, TT,TE,EE+lowE), (12) corresponding to an anti-correlation between the matter density Ωmh2 and the Hubble p",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/planck2018-cosmology.html",
      "source_body_sha256": "3d9ded3ff4f8608bf90fa832d09d397b49eb0a1e3c0de1ff46867c52e90a2b73",
      "evidence_sha256": "1c5b0d9f071c5b1efcbcc6c3ce5a836faf10c09113dcace5a6c408e97f9685e2",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "degrees-of-freedom",
     "evidence": "Values for χ2 p are given in Table 3 together with prob- abilities to exceed (PTEs) computed from a χ2 distribution with five degrees of freedom.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "degrees of freedom in a statistical model/test",
     "provenance": {
      "source_path": "papers/planck2018-cosmology.html",
      "source_body_sha256": "3d9ded3ff4f8608bf90fa832d09d397b49eb0a1e3c0de1ff46867c52e90a2b73",
      "evidence_sha256": "19b3bb83ecb46c47c1713d71d34cfcc3ad6c779d129dcc4de2807fd2247c35e9",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "mean",
     "evidence": "ift bins may depend on photometric redshift uncertainties that are not well captured by a single shift in the mean of each window’s population.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/planck2018-cosmology.html",
      "source_body_sha256": "3d9ded3ff4f8608bf90fa832d09d397b49eb0a1e3c0de1ff46867c52e90a2b73",
      "evidence_sha256": "ba7b9356e085811011ddab1801d41d6e3bc180f2320f9fb66bab02cf91fb1816",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "null-hypothesis",
     "evidence": "Assuming the base-ΛCDM cosmology, the inferred (model-dependent) late-Universe parameters are: Hubble constant H0 = (67.4±0.5) km s−1Mpc−1; matter density parameter Ωm = 0.315±0.00",
     "decision": "rejected",
     "role": "symbol-collision",
     "reason": "H0 denotes a Hubble parameter, Hamiltonian, operator, or generating function rather than a null hypothesis",
     "provenance": {
      "source_path": "papers/planck2018-cosmology.html",
      "source_body_sha256": "3d9ded3ff4f8608bf90fa832d09d397b49eb0a1e3c0de1ff46867c52e90a2b73",
      "evidence_sha256": "bb96b426a0705b577e0b884a03fadf466352e0e2b29bef46d2c2461948a010a2",
      "grounding_status": "not-exactly-located"
     }
    },
    {
     "term": "p-value",
     "evidence": "PArthENoPE predicts Y BBN P = 0.24672 +(0.00011)0.00061 −(0.00012)0.00061 (95 %, TT,TE,EE+lowE), (72a) while PRIMAT gives Y BBN P = 0.24714 +(0.00012)0.00049 −(0.00013)0.00049 (95 ",
     "decision": "rejected",
     "role": "non-statistical-sense",
     "reason": "lexical match does not denote the controlled statistical concept",
     "provenance": {
      "source_path": "papers/planck2018-cosmology.html",
      "source_body_sha256": "3d9ded3ff4f8608bf90fa832d09d397b49eb0a1e3c0de1ff46867c52e90a2b73",
      "evidence_sha256": "d2fcf6004675fe87eddeb020725f1c1426cbd9be0515c146b65d1f2aca7d2209",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "standard-deviation",
     "evidence": "Results from our main grid of parameter constraints on one- or two-parameter extensions to the base-ΛCDM cosmol- 2Here σ2013 is the standard deviation quoted on parameters in PCP13",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/planck2018-cosmology.html",
      "source_body_sha256": "3d9ded3ff4f8608bf90fa832d09d397b49eb0a1e3c0de1ff46867c52e90a2b73",
      "evidence_sha256": "62525b58461b964f08d6f50994786bb5038f020229e863fe3c5b5234181c529e",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "statistical-power",
     "evidence": "2 σ) on the base-ΛCDM results when combining into the full “TT,TE,EE,” Plik likelihood be- cause EE has lower statistical power compared to T T or T E; however, dust modelling in E",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/planck2018-cosmology.html",
      "source_body_sha256": "3d9ded3ff4f8608bf90fa832d09d397b49eb0a1e3c0de1ff46867c52e90a2b73",
      "evidence_sha256": "15be22d0066ecb294ab23304d6a9096f0122c342f38b9a854c4d2305958596d9",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "statistical-significance",
     "evidence": "2.3) at moderate statistical significance, raising the question of whether there are unaccounted for systematic effects lurking within the Planck data.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit inferential-statistical context",
     "provenance": {
      "source_path": "papers/planck2018-cosmology.html",
      "source_body_sha256": "3d9ded3ff4f8608bf90fa832d09d397b49eb0a1e3c0de1ff46867c52e90a2b73",
      "evidence_sha256": "38f8ec5e75d1835dff9196115292aadfa2caeed12beac8f46bbf0274debedefd",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "variance",
     "evidence": "The error bars show ±1 σ diagonal uncertainties, including cosmic variance (approximated as Gaussian) and not including uncertainties in the foreground model at (cid:96) ≥ 30.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/planck2018-cosmology.html",
      "source_body_sha256": "3d9ded3ff4f8608bf90fa832d09d397b49eb0a1e3c0de1ff46867c52e90a2b73",
      "evidence_sha256": "3b0cddc936fe0d3d600262e4f1be10989c9b38cab286af426082be54c0a9fb40",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "arXiv-nonexclusive-distrib-1.0",
   "license_url": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
   "source_url": "https://arxiv.org/abs/1807.06209",
   "doi": "10.1051/0004-6361/201833910",
   "full_text_redistributable": false
  },
  {
   "id": "niv2009-rl-brain",
   "title": "Reinforcement Learning in the Brain",
   "discipline": "neuroscience",
   "year": 2009,
   "authors": "Yael Niv",
   "url": "papers/niv2009-rl-brain.html",
   "key_terms": [
    "reinforcement learning",
    "reward prediction error",
    "dopamine",
    "decision-making",
    "conditioning",
    "normative framework",
    "temporal difference",
    "action selection",
    "basal ganglia",
    "instrumental conditioning"
   ],
   "phenomenon": "reinforcement learning in the brain (temporal-difference reward prediction error / dopamine)",
   "claim_type": "existence",
   "method": "theory",
   "statistics": [
    {
     "term": "bootstrap",
     "evidence": "he soundness and convergence of ‘policy iteration’ methods (in which a policy is iteratively improved through bootstrapping of the values derived given each policy; Howard, 1960; B",
     "decision": "rejected",
     "role": "non-statistical-sense",
     "reason": "lexical match does not denote the controlled statistical concept",
     "provenance": {
      "source_path": "papers/niv2009-rl-brain.html",
      "source_body_sha256": "22a6177fbc684a974f20e3fcdd363e621b599c3e800987d033b77e342a3a4865",
      "evidence_sha256": "1aefb299a2f71ac8cee809e29ad74fc0b8b255d349b9f4ee39fe1049cbfb7baa",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "mean",
     "evidence": "rewards to the dopaminergic response to the current reward is exactly according to an exponentially weighted average of past experience, as is implied by the TD learning rule (Baye",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/niv2009-rl-brain.html",
      "source_body_sha256": "22a6177fbc684a974f20e3fcdd363e621b599c3e800987d033b77e342a3a4865",
      "evidence_sha256": "a735a7231411f962d2077146172c04c5aca97a5fd05867d48ac08dd4766d2dbd",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "p-value",
     "evidence": "Moreover, conditioned stimuli predicting probabilistic rewards or rewards 14 abc010101015152501p=0 p=0.25 p=0.5 p=0.75 p=1 stimulusonrewardor stimulus offstimulusonrewardor stimulu",
     "decision": "rejected",
     "role": "non-statistical-sense",
     "reason": "lexical match does not denote the controlled statistical concept",
     "provenance": {
      "source_path": "papers/niv2009-rl-brain.html",
      "source_body_sha256": "22a6177fbc684a974f20e3fcdd363e621b599c3e800987d033b77e342a3a4865",
      "evidence_sha256": "aa632d383a040cf2151ca306e5d9b5508e2ae6c63c38633b61410713cd412c27",
      "grounding_status": "not-exactly-located"
     }
    },
    {
     "term": "regression",
     "evidence": "For instance, using general linear regression, Bayer and 13 Figure 3: Dopaminergic firing patterns comply with the predictions of TD learning.",
     "decision": "accepted",
     "role": "substantive-mention",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/niv2009-rl-brain.html",
      "source_body_sha256": "22a6177fbc684a974f20e3fcdd363e621b599c3e800987d033b77e342a3a4865",
      "evidence_sha256": "196c1ee592accf8c96fa927764264957079bade725b5b5c4c64599c15c6b9b94",
      "grounding_status": "not-exactly-located"
     }
    },
    {
     "term": "standard-deviation",
     "evidence": "Behavioral results suggest that interval timing is ex- tremely inaccurate, with the standard deviation of the prediction of an interval being proportional to the mean length of the",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/niv2009-rl-brain.html",
      "source_body_sha256": "22a6177fbc684a974f20e3fcdd363e621b599c3e800987d033b77e342a3a4865",
      "evidence_sha256": "7a97b0e732e261c24d77d436284dda959bad70d0daa8b1453caeff5a2133ebfe",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "open-access",
   "license_url": "",
   "source_url": "https://www.princeton.edu/~yael/Publications/Niv2009.pdf",
   "doi": "10.1016/j.jmp.2008.12.005",
   "full_text_redistributable": false
  },
  {
   "id": "gu2020-ml-asset-pricing",
   "title": "Empirical Asset Pricing via Machine Learning",
   "discipline": "quantitative-finance",
   "year": 2020,
   "authors": "Shihao Gu, Bryan Kelly, Dacheng Xiu",
   "url": "papers/gu2020-ml-asset-pricing.html",
   "key_terms": [],
   "phenomenon": "",
   "claim_type": "",
   "method": "computational-model",
   "statistics": [
    {
     "term": "Bayesian",
     "evidence": "First, they use a modified ridge penalty based on a Bayesian prior.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/gu2020-ml-asset-pricing.html",
      "source_body_sha256": null,
      "evidence_sha256": "a1050af05bdaad6d0a5091f4a71bae2d94c9468a150d085858751d010f466c7c",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "R-squared",
     "evidence": "s that all approaches are exposed to similar transaction costs and it is valid to directly compare their risk-adjusted returns.44 Avramov, Cheng, and Metzker (2020) raise the conce",
     "decision": "rejected",
     "role": "non-statistical-sense",
     "reason": "lexical match does not denote the controlled statistical concept",
     "provenance": {
      "source_path": "papers/gu2020-ml-asset-pricing.html",
      "source_body_sha256": null,
      "evidence_sha256": "e4d750383391d2eb4f9e27268550df1eba855bb9a92d3f83c59da12820543fb2",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "confidence-interval",
     "evidence": "However, this does not affect the estimation procedure and our discussion, but only the confidence intervals.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/gu2020-ml-asset-pricing.html",
      "source_body_sha256": null,
      "evidence_sha256": "50061c1d604936a73b08ff9aef7683b2cda1b6a7ac6a803e76af8bb698c10c3e",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "correlation",
     "evidence": "SDF Structure Table A.V: Correlation of GAN SDF with Fama-French 5 Factors and Time-Series Regression Mkt-RF SMB HML RMW CMA intercept Regression Coefficients Correlations 0.00 (0.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/gu2020-ml-asset-pricing.html",
      "source_body_sha256": null,
      "evidence_sha256": "dc609acd0c08f5fd94a5e063da7b6391d2ba61310c4645eb806ea7262bf4d9d5",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "hypothesis-testing",
     "evidence": "The GRS-test is under the null hypothesis of correctly pricing all decile portfolios and includes the p-values.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit hypothesis-testing language",
     "provenance": {
      "source_path": "papers/gu2020-ml-asset-pricing.html",
      "source_body_sha256": null,
      "evidence_sha256": "aaaddaef35374144edfd789e33462afc5bca00d44e6c8c44a023a01c22af2b94",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "mean",
     "evidence": "However, the risk premia of individual stocks is time-varying and an unconditional mean of stock returns might not capture the predictable component.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "numeric mean or average in main text",
     "provenance": {
      "source_path": "papers/gu2020-ml-asset-pricing.html",
      "source_body_sha256": null,
      "evidence_sha256": "d37632d59c8c5af6b411831c2a7bcdde6c8823a9cad6236886d3c95bebe0830e",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "median",
     "evidence": ".20.00.20.4BEME0.40.20.00.20.4ST_REV0.40.20.00.20.40.051910.039170.026430.013690.000950.011800.024540.03728weight characteristic that can be different than the median.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/gu2020-ml-asset-pricing.html",
      "source_body_sha256": null,
      "evidence_sha256": "bd6a68ba68b3876c0516ed4a18ef04143cdfa0db9444103ecdb21c668dfe4338",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "null-hypothesis",
     "evidence": "The GRS-test is under the null hypothesis of correctly pricing all decile portfolios and includes the p-values.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit null-hypothesis language",
     "provenance": {
      "source_path": "papers/gu2020-ml-asset-pricing.html",
      "source_body_sha256": null,
      "evidence_sha256": "aaaddaef35374144edfd789e33462afc5bca00d44e6c8c44a023a01c22af2b94",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "p-value",
     "evidence": "The GRS-test is under the null hypothesis of correctly pricing all decile portfolios and includes the p-values.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit p-value or inferential p notation",
     "provenance": {
      "source_path": "papers/gu2020-ml-asset-pricing.html",
      "source_body_sha256": null,
      "evidence_sha256": "aaaddaef35374144edfd789e33462afc5bca00d44e6c8c44a023a01c22af2b94",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "regression",
     "evidence": "The risk exposure ˆβLS is obtained by a linear regression of Re t+1Ft+1 on It,i.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/gu2020-ml-asset-pricing.html",
      "source_body_sha256": null,
      "evidence_sha256": "aca203e71d49019fa5d5f189a20a69812198c651a52841009978e4f5ca23d3fb",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "sample-size",
     "evidence": "The panel data set is N = 500, T = 600, where the first Ttrain = 250 are used for training, the next Tvalid = 100 observations are the validation and the last Ttest = 250 observati",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/gu2020-ml-asset-pricing.html",
      "source_body_sha256": null,
      "evidence_sha256": "c4b2ddd4e099fe0312eee5d02e772a3f94bb34b8b23659c304665b6549cd21b5",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "standard-deviation",
     "evidence": "Figure 16 plots the cumulative return for each model normalized by the standard deviation.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/gu2020-ml-asset-pricing.html",
      "source_body_sha256": null,
      "evidence_sha256": "53bfb0447165b63ab3baaf8248420f1494936934e4a642e0afbad13af0deafef",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "standard-error",
     "evidence": "Standard errors are in parenthesis.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit standard error or SEM",
     "provenance": {
      "source_path": "papers/gu2020-ml-asset-pricing.html",
      "source_body_sha256": null,
      "evidence_sha256": "bc5a60cde88285faa2197e83aa3832d12de4577f2ffaa254c7e5831d1b0b7621",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "t-test",
     "evidence": "GRS 10.75 -4.35 -2.68 -1.05 -0.54 2.17 2.20 3.41 4.86 5.39 6.59 9.96 p 0.00 This table shows the average returns, time series pricing errors and corresponding t-statistics for β-so",
     "decision": "rejected",
     "role": "non-statistical-sense",
     "reason": "lexical match does not denote the controlled statistical concept",
     "provenance": {
      "source_path": "papers/gu2020-ml-asset-pricing.html",
      "source_body_sha256": null,
      "evidence_sha256": "1f26a5774328d8b9b48653d8a98c0bb048daf25e45652e83efbc6d24615d9784",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "variance",
     "evidence": "Following Cochrane (2003) we use the above normalized relationship between the SDF and the mean-variance efficient portfolio.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/gu2020-ml-asset-pricing.html",
      "source_body_sha256": null,
      "evidence_sha256": "1b8caec7fc4dce1e54f50dab4775222ee9c115bf17aec334048d4074044436da",
      "grounding_status": "source-unavailable"
     }
    }
   ],
   "license_spdx": "arxiv-open",
   "source_url": "https://arxiv.org/abs/1904.00745",
   "doi": "10.1093/rfs/hhaa009",
   "full_text_redistributable": false
  },
  {
   "id": "schulz2010-consort",
   "title": "CONSORT 2010 Statement: Updated Guidelines for Reporting Parallel Group Randomised Trials",
   "discipline": "medicine",
   "year": 2010,
   "authors": "Kenneth F. Schulz, Douglas G. Altman, David Moher",
   "url": "papers/schulz2010-consort.html",
   "key_terms": [
    "randomised controlled trials",
    "CONSORT",
    "checklist",
    "flow diagram",
    "allocation concealment",
    "blinding",
    "transparent reporting",
    "randomisation",
    "trial registration",
    "methodological rigour"
   ],
   "phenomenon": "transparent reporting standards for randomized trials (reducing bias)",
   "claim_type": "method",
   "method": "survey",
   "statistics": [
    {
     "term": "confidence-interval",
     "evidence": "y and secondary outcome, results for each group, and the estimated effect size and its precision (such as 95% confidence interval) For binary outcomes, presentation of both absolut",
     "decision": "accepted",
     "role": "substantive-mention",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/schulz2010-consort.html",
      "source_body_sha256": "d948b46df398e69ccde7d7e1471d996901f19678145511aae435b9a4ba8ec486",
      "evidence_sha256": "0944eafaa9886c7fecb929bba93c5492c39a67e772eda52c0eda28e423889e79",
      "grounding_status": "not-exactly-located"
     }
    },
    {
     "term": "effect-size",
     "evidence": "by original assigned groups For each primary and secondary outcome, results for each group, and the estimated effect size and its precision (such as 95% confidence interval) For bi",
     "decision": "accepted",
     "role": "substantive-mention",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/schulz2010-consort.html",
      "source_body_sha256": "d948b46df398e69ccde7d7e1471d996901f19678145511aae435b9a4ba8ec486",
      "evidence_sha256": "36cc4f89e3eccaed0b321dc217b2b841a8db40635a37904b4803a1838e5fa2e4",
      "grounding_status": "not-exactly-located"
     }
    },
    {
     "term": "sample-size",
     "evidence": "ing how and when they were assessed Any changes to trial outcomes after the trial commenced, with reasons How sample size was determined When applicable, explanation of any interim",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/schulz2010-consort.html",
      "source_body_sha256": "d948b46df398e69ccde7d7e1471d996901f19678145511aae435b9a4ba8ec486",
      "evidence_sha256": "8470cc9fbcb07cbc12d63aaea98597b65ed902fc0864085d960bc3e7673cec24",
      "grounding_status": "not-exactly-located"
     }
    }
   ],
   "license_spdx": "CC-BY-4.0",
   "license_url": "http://creativecommons.org/licenses/by/4.0/",
   "source_url": "https://journals.plos.org/plosmedicine/article?id=10.1371/journal.pmed.1000251",
   "doi": "10.1371/journal.pmed.1000251",
   "full_text_redistributable": true
  },
  {
   "id": "garimella2017-polarization",
   "title": "A Long-Term Analysis of Polarization on Twitter",
   "discipline": "political-science",
   "year": 2017,
   "authors": "Kiran Garimella, Ingmar Weber",
   "url": "papers/garimella2017-polarization.html",
   "key_terms": [
    "polarization",
    "Twitter",
    "users",
    "political",
    "retweet",
    "hashtags",
    "network",
    "partisan",
    "content"
   ],
   "phenomenon": "political polarization on social media (echo chambers)",
   "claim_type": "existence",
   "method": "computational-model",
   "statistics": [
    {
     "term": "Bayesian",
     "evidence": "com/Jefferson-Henrique/GetOldTweets-python for polarization, we decided to apply a Bayesian methodol- ogy.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/garimella2017-polarization.html",
      "source_body_sha256": "ef5be0ec3ebd21e12035aea83a5a3fc428fa07f030e806a12f9da91acb5270cc",
      "evidence_sha256": "5938ccc2a5a2a99ded72efcd3b175431b65c40e7881fb1adc7cc25e57d8fe652",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "mean",
     "evidence": "The mean of the beta distribu- tion, and hence the “leaning” l of the following (retweeting) user, is defined as l = α/(α + β), taking the leftness as the direction of the index.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "numeric mean or average in main text",
     "provenance": {
      "source_path": "papers/garimella2017-polarization.html",
      "source_body_sha256": "ef5be0ec3ebd21e12035aea83a5a3fc428fa07f030e806a12f9da91acb5270cc",
      "evidence_sha256": "2268b249c568e97e2615b10fb74b7d69fc86c88a0252c2b8df56eb9fcc931258",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "p-value",
     "evidence": "An augmented Dickey-Fuller test found the time series to be non-stationary (p < 0.0001).",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit p-value or inferential p notation",
     "provenance": {
      "source_path": "papers/garimella2017-polarization.html",
      "source_body_sha256": "ef5be0ec3ebd21e12035aea83a5a3fc428fa07f030e806a12f9da91acb5270cc",
      "evidence_sha256": "abb5d4d7f06f0c17deab9454d51de1b993d3e282afaf1a4fea459919bf4fb225",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "t-test",
     "evidence": "Next, we computed the linear fit across time and tested the value of the non-zero slope for statistical sig- nificance using a t-test (p < 0.0001).",
     "decision": "accepted",
     "role": "substantive-mention",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/garimella2017-polarization.html",
      "source_body_sha256": "ef5be0ec3ebd21e12035aea83a5a3fc428fa07f030e806a12f9da91acb5270cc",
      "evidence_sha256": "6acc31455739ea4be66f85d160ba6ca641be03f8fd0f6c5e343bcac51d7ae4fb",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "arXiv-nonexclusive-distrib-1.0",
   "license_url": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
   "source_url": "https://arxiv.org/abs/1703.02769",
   "doi": "",
   "full_text_redistributable": false
  },
  {
   "id": "sole2001-ecological-networks",
   "title": "Complexity and Fragility in Ecological Networks",
   "discipline": "ecology",
   "year": 2001,
   "authors": "Ricard V. Solé, José M. Montoya",
   "url": "papers/sole2001-ecological-networks.html",
   "key_terms": [
    "network",
    "species",
    "scale-free",
    "robust",
    "fragile",
    "extinction",
    "nodes",
    "connections",
    "trophic"
   ],
   "phenomenon": "scale-free network (ecological food webs; robust-yet-fragile)",
   "claim_type": "existence",
   "method": "simulation",
   "statistics": [
    {
     "term": "p-value",
     "evidence": "The estimated slopes are indicated ((a) r2 = 0.87; p < 0.01; (b) r2 = 0.87; p < 0.01; (c) r2 = 0.92; p < 0.01; (d) r2 = 0.77; p < 0.01.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit p-value or inferential p notation",
     "provenance": {
      "source_path": "papers/sole2001-ecological-networks.html",
      "source_body_sha256": "9c807a988f803e69c69410fc10d4d67ac430a415d15bcb8ffa0e463cb1c91dac",
      "evidence_sha256": "ecbbf0163457435f326ef430966ac42ec8e8bcfb098a4fa97e53b5aefbc94ab4",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "arxiv-open",
   "license_url": "",
   "source_url": "https://arxiv.org/abs/cond-mat/0011196",
   "doi": "10.1098/rspb.2001.1767",
   "full_text_redistributable": false
  },
  {
   "id": "candes2006-robust-uncertainty",
   "title": "Robust Uncertainty Principles: Exact Signal Reconstruction from Highly Incomplete Frequency Information",
   "discipline": "information-theory",
   "year": 2006,
   "authors": "Emmanuel J. Candès, Justin Romberg, Terence Tao",
   "url": "papers/candes2006-robust-uncertainty.html",
   "key_terms": [
    "signal",
    "Fourier",
    "sparse",
    "reconstruction",
    "frequency",
    "minimization",
    "convex",
    "coefficients",
    "recovery"
   ],
   "phenomenon": "compressed sensing (exact signal recovery from highly incomplete measurements)",
   "claim_type": "existence",
   "method": "analytic",
   "statistics": [
    {
     "term": "null-hypothesis",
     "evidence": "15 Theorem 3.1 Put H0 = ι∗H for short, where H is the operator defined by (2.10).",
     "decision": "rejected",
     "role": "symbol-collision",
     "reason": "H0 denotes a Hubble parameter, Hamiltonian, operator, or generating function rather than a null hypothesis",
     "provenance": {
      "source_path": "papers/candes2006-robust-uncertainty.html",
      "source_body_sha256": "0e48ec1d6c849da36efdae723d8a25d93f9db937467d6f393e6d106af265080b",
      "evidence_sha256": "d01975d1696f86f6a384b95592fb341062d1fc24d554cd06ac59180d73eead0e",
      "grounding_status": "not-exactly-located"
     }
    },
    {
     "term": "regression",
     "evidence": "Indeed with knowledge of T , a reasonable strategy might be to recover f by the method of least-squares, namely, → In practice, the matrix inversion might be problematic.",
     "decision": "rejected",
     "role": "non-statistical-sense",
     "reason": "lexical match does not denote the controlled statistical concept",
     "provenance": {
      "source_path": "papers/candes2006-robust-uncertainty.html",
      "source_body_sha256": "0e48ec1d6c849da36efdae723d8a25d93f9db937467d6f393e6d106af265080b",
      "evidence_sha256": "6dd3b478463a09eb2746c97482755d5fa502bec0075eb5fff8d2b49fbed655cf",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "sample-size",
     "evidence": "Note that the problem here does not really have anything to do with l1-minimization per se; f cannot be reconstructed from its Fourier samples on Ω thereby showing that Theorem 1.1",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/candes2006-robust-uncertainty.html",
      "source_body_sha256": "0e48ec1d6c849da36efdae723d8a25d93f9db937467d6f393e6d106af265080b",
      "evidence_sha256": "c2d75acdd36e14d90ddee2182382f0865b8833362011a86b6d6005de04982c22",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "variance",
     "evidence": "Randomly generate f by setting f (t) = 0, t T c and drawing both the real and T from independent Gaussian distributions with mean imaginary parts of f (t), t zero and variance one2",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/candes2006-robust-uncertainty.html",
      "source_body_sha256": "0e48ec1d6c849da36efdae723d8a25d93f9db937467d6f393e6d106af265080b",
      "evidence_sha256": "3d1a69e1cfaa2f985170ea717e4f8ea0b1b64782f1b9ac5d9f9ec39c904c0087",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "arxiv-open",
   "license_url": "",
   "source_url": "https://arxiv.org/abs/math/0409186",
   "doi": "10.1109/TIT.2005.862083",
   "full_text_redistributable": false
  },
  {
   "id": "pastorsatorras2001-epidemic-scalefree",
   "title": "Epidemic Spreading in Scale-Free Networks",
   "discipline": "epidemiology",
   "year": 2001,
   "authors": "Romualdo Pastor-Satorras, Alessandro Vespignani",
   "url": "papers/pastorsatorras2001-epidemic-scalefree.html",
   "key_terms": [
    "epidemic",
    "network",
    "scale-free",
    "spreading",
    "threshold",
    "virus",
    "power-law",
    "infection",
    "connectivity"
   ],
   "phenomenon": "epidemic spreading on scale-free networks (absence of an epidemic threshold)",
   "claim_type": "mechanism",
   "method": "analytic",
   "statistics": [
    {
     "term": "statistical-significance",
     "evidence": "The scale-free connectivity of the Internet implies that each node has a statistically significant probability of having a very large number of connections compared to the average ",
     "decision": "rejected",
     "role": "lexical-significance",
     "reason": "significant/significantly is used without inferential-statistical meaning",
     "provenance": {
      "source_path": "papers/pastorsatorras2001-epidemic-scalefree.html",
      "source_body_sha256": "8a91146ae86dba94e6b820ad83c369123dd560c4899c8ee34d948c991d804f40",
      "evidence_sha256": "e3d619c14e9cb3bbbbad14d4ec714737b74500dd4e67bf87d8215d8743d1f1f0",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "arxiv-open",
   "license_url": "",
   "source_url": "https://arxiv.org/abs/cond-mat/0010317",
   "doi": "10.1103/PhysRevLett.86.3200",
   "full_text_redistributable": false
  },
  {
   "id": "gatheral2018-rough-volatility",
   "title": "Volatility Is Rough",
   "discipline": "quantitative-finance",
   "year": 2018,
   "authors": "Jim Gatheral, Thibault Jaisson, Mathieu Rosenbaum",
   "url": "papers/gatheral2018-rough-volatility.html",
   "key_terms": [
    "volatility",
    "fractional",
    "Brownian",
    "Hurst",
    "log-volatility",
    "exponent",
    "memory",
    "realized",
    "rough"
   ],
   "phenomenon": "rough volatility (monofractal / fractional-Brownian scaling of financial volatility)",
   "claim_type": "existence",
   "method": "analytic",
   "statistics": [
    {
     "term": "correlation",
     "evidence": "Finally, when H > 1/2, the increments of the fBM are positively correlated and exhibit long memory in the sense that +∞ (cid:88) Cov[W H 1 , W H k − W H k−1] = +∞.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/gatheral2018-rough-volatility.html",
      "source_body_sha256": "42d1d3bc1540160b3d6001b80b12d5879412c533a021887a975a1e36e3483167",
      "evidence_sha256": "292cd77d59a6755a7c3300a18b774c56ec2ae9e4f6830e282583382a73de3a44",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "mean",
     "evidence": "t+∆)] denotes the empirical mean of the log-variance over the SPX2.rv ∆ = 1 SPX2.rv ∆ = 5 SPX2.rv ∆ = 20 FTSE2.rv ∆ = 1 FTSE2.rv ∆ = 5 FTSE2.rv ∆ = 20 N2252.rv ∆ = 1 N2252.rv ∆ = 5",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/gatheral2018-rough-volatility.html",
      "source_body_sha256": "42d1d3bc1540160b3d6001b80b12d5879412c533a021887a975a1e36e3483167",
      "evidence_sha256": "8ab33e2465eaa32f954d532da257d5759ff77edd1f18ffe625c912a518f3c0bf",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "regression",
     "evidence": "For each index and for q = 0.5, 1, 1.5, 2, 3, by doing a linear regression of log(m(q, ∆)) on log(∆) for ∆ = 1, ..., 30, we obtain estimates of ζq that we summarize in Table B.1 in",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/gatheral2018-rough-volatility.html",
      "source_body_sha256": "42d1d3bc1540160b3d6001b80b12d5879412c533a021887a975a1e36e3483167",
      "evidence_sha256": "10fa0971318f87e92b9edffd43d37260292beac813091628186db5b3394839c7",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "sample-size",
     "evidence": "We conclude that this effect relates to finite sample size and is thus not signifi- cant.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/gatheral2018-rough-volatility.html",
      "source_body_sha256": "42d1d3bc1540160b3d6001b80b12d5879412c533a021887a975a1e36e3483167",
      "evidence_sha256": "563cb05a25aecb42b09e19f91a2252ed52c840108a5fc61d2e166456147537c4",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "standard-error",
     "evidence": "Figure 4.2: Autocorrelation functions of log(σt) (in blue) and εt (in green) and the Bartlett standard error bands (in red), for S&P data (above) and for simulated data (below).",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit standard error or SEM",
     "provenance": {
      "source_path": "papers/gatheral2018-rough-volatility.html",
      "source_body_sha256": "42d1d3bc1540160b3d6001b80b12d5879412c533a021887a975a1e36e3483167",
      "evidence_sha256": "402a86bf9bd4072c3efa2b9628b80ab8126ab522b2b68cded18aff03013013a1",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "statistical-significance",
     "evidence": "Statistically significant fits of Hawkes processes to financial data.",
     "decision": "rejected",
     "role": "bibliography",
     "reason": "occurrence is confined to a citation/reference rather than the article body",
     "provenance": {
      "source_path": "papers/gatheral2018-rough-volatility.html",
      "source_body_sha256": "42d1d3bc1540160b3d6001b80b12d5879412c533a021887a975a1e36e3483167",
      "evidence_sha256": "29662984854136b3e2e6e5898c6a86727b84b896dd8473c4861756386af9dd01",
      "grounding_status": "verified-reference"
     }
    },
    {
     "term": "variance",
     "evidence": "the volatility process for four assets: The DAX and Bund futures contracts, for which we estimate integrated variance directly from high frequency data using an estimator based on ",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/gatheral2018-rough-volatility.html",
      "source_body_sha256": "42d1d3bc1540160b3d6001b80b12d5879412c533a021887a975a1e36e3483167",
      "evidence_sha256": "c5c54a331291be5bcf93fc2175075c872d5b82dbbd369292e2968ad322af3f40",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "arXiv-nonexclusive-distrib-1.0",
   "license_url": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
   "source_url": "https://arxiv.org/abs/1410.3394",
   "doi": "10.1080/14697688.2017.1393551",
   "full_text_redistributable": false
  },
  {
   "id": "ongaro2014-raft",
   "title": "In Search of an Understandable Consensus Algorithm",
   "discipline": "computer-science",
   "year": 2014,
   "authors": "Diego Ongaro, John Ousterhout",
   "url": "papers/ongaro2014-raft.html",
   "key_terms": [
    "consensus",
    "Raft",
    "Paxos",
    "replication",
    "leader",
    "safety",
    "log",
    "understandable",
    "majority"
   ],
   "phenomenon": "distributed consensus (Raft; understandable consensus algorithm for a replicated log)",
   "claim_type": "method",
   "method": "design",
   "statistics": [
    {
     "term": "regression",
     "evidence": "We also created a linear regression model that predicts a new student’s quiz scores based on three factors: which quiz they took, their degree of prior Paxos experience, and s t n ",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/ongaro2014-raft.html",
      "source_body_sha256": "6f9084af8c1f3114f8b16e533b360cd12397d2a3dec87f2eb2f4990ccee2fe31",
      "evidence_sha256": "297c10986075fd90462a53089c558488ac637efc6d03525440b326f12e7ec32e",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "statistical-significance",
     "evidence": "Each implementation begins with Paxos, dis- covers the difficulties in implementing it, and then de- velops a significantly different architecture.",
     "decision": "rejected",
     "role": "lexical-significance",
     "reason": "significant/significantly is used without inferential-statistical meaning",
     "provenance": {
      "source_path": "papers/ongaro2014-raft.html",
      "source_body_sha256": "6f9084af8c1f3114f8b16e533b360cd12397d2a3dec87f2eb2f4990ccee2fe31",
      "evidence_sha256": "e3aa37c80f34236b555672ceb6714fe2e522a0ef4beb80f6a3c667955158f35e",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "t-test",
     "evidence": "A paired t-test states that, with 95% confidence, the true distribution of Raft scores has a mean at least 2.5 points larger than the true distribution of Paxos scores.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/ongaro2014-raft.html",
      "source_body_sha256": "6f9084af8c1f3114f8b16e533b360cd12397d2a3dec87f2eb2f4990ccee2fe31",
      "evidence_sha256": "723cfb94fd034f5e2ace4ae3afe201ad897332886564659f0e6fd615f117c824",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "author-open",
   "license_url": "",
   "source_url": "https://raft.github.io/raft.pdf",
   "doi": "",
   "full_text_redistributable": false
  },
  {
   "id": "bojarski2016-self-driving",
   "title": "End to End Learning for Self-Driving Cars",
   "discipline": "robotics",
   "year": 2016,
   "authors": "Mariusz Bojarski, Davide Del Testa, Daniel Dworakowski, Bernhard Firner",
   "url": "papers/bojarski2016-self-driving.html",
   "key_terms": [
    "convolutional",
    "steering",
    "training",
    "network",
    "camera",
    "road",
    "image",
    "autonomy",
    "features"
   ],
   "phenomenon": "end-to-end learning for autonomous driving (pixels to steering)",
   "claim_type": "method",
   "method": "experiment",
   "statistics": [
    {
     "term": "standard-deviation",
     "evidence": "The distribution has zero mean, and the standard deviation is twice the standard deviation that we measured with human drivers.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/bojarski2016-self-driving.html",
      "source_body_sha256": "302adfe3074842198ea85db5cefe9c424692f5308ae6d7d1578e2d26efaf2e62",
      "evidence_sha256": "b0dab0809e33fc9d204592180c46a4c69f1f2a287c3d3fd7664ab7f8358c8bab",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "arXiv-nonexclusive-distrib-1.0",
   "license_url": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
   "source_url": "https://arxiv.org/abs/1604.07316",
   "doi": "",
   "full_text_redistributable": false
  },
  {
   "id": "hansen2013-climate-change",
   "title": "Assessing Dangerous Climate Change: Required Reduction of Carbon Emissions to Protect Young People, Future Generations and Nature",
   "discipline": "earth-science",
   "year": 2013,
   "authors": "James Hansen, Pushker Kharecha, Makiko Sato",
   "url": "papers/hansen2013-climate-change.html",
   "key_terms": [
    "climate change",
    "carbon emissions",
    "fossil fuel",
    "global warming",
    "paleoclimate",
    "energy imbalance",
    "carbon cycle",
    "feedbacks",
    "CO2",
    "emissions"
   ],
   "phenomenon": "dangerous anthropogenic climate change (carbon emission targets)",
   "claim_type": "critique",
   "method": "analytic",
   "statistics": [
    {
     "term": "correlation",
     "evidence": "Annual changes of CO2 are highly correlated with the El Nin ̃ o cycle (Fig.",
     "decision": "accepted",
     "role": "substantive-mention",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/hansen2013-climate-change.html",
      "source_body_sha256": "a5cc1cf32924613068099d5549ed031e1e5f89ab787fddbe1c68e37fd0150d96",
      "evidence_sha256": "b4a1821f97c4d6484fde90e88b7b8250b89f0901f9a7b1c5dc580850d02dc7f1",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "mean",
     "evidence": "rent prices) and resources (potentially recoverable with advanced technology and/or at higher prices) are the mean of estimates of Energy Information Administration (EIA) [7], Germ",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/hansen2013-climate-change.html",
      "source_body_sha256": "a5cc1cf32924613068099d5549ed031e1e5f89ab787fddbe1c68e37fd0150d96",
      "evidence_sha256": "a9f86c71b31d5a5e931c22026dd7c2a17a71382b4d4a9285e4c04a18022cc51e",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "median",
     "evidence": "A probabilistic analysis [20], still partly subjective, found a median ‘‘dangerous’’ threshold of 2.8uC, with 95% confidence that threshold was 1.5uC or higher.",
     "decision": "accepted",
     "role": "substantive-mention",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/hansen2013-climate-change.html",
      "source_body_sha256": "a5cc1cf32924613068099d5549ed031e1e5f89ab787fddbe1c68e37fd0150d96",
      "evidence_sha256": "7fcb15a9e74bfcca652164e9d85f29e0d886a2ba0deedc5136f1cc621b33f761",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "statistical-significance",
     "evidence": "7) support the conclusion that the solar irradiance in the current solar cycle is significantly lower than in the three preceding solar cycles.",
     "decision": "rejected",
     "role": "lexical-significance",
     "reason": "significant/significantly is used without inferential-statistical meaning",
     "provenance": {
      "source_path": "papers/hansen2013-climate-change.html",
      "source_body_sha256": "a5cc1cf32924613068099d5549ed031e1e5f89ab787fddbe1c68e37fd0150d96",
      "evidence_sha256": "1eb5e36659b5cedced0b2013eb58851039f9d350a8b5fdcdf15e83379f935cb6",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "CC0-1.0",
   "license_url": "http://creativecommons.org/publicdomain/zero/1.0/",
   "source_url": "https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0081648",
   "doi": "10.1371/journal.pone.0081648",
   "full_text_redistributable": true
  },
  {
   "id": "kozlowski2019-geometry-culture",
   "title": "The Geometry of Culture: Analyzing the Meanings of Class through Word Embeddings",
   "discipline": "digital-humanities",
   "year": 2019,
   "authors": "Austin C. Kozlowski, Matt Taddy, James A. Evans",
   "url": "papers/kozlowski2019-geometry-culture.html",
   "key_terms": [
    "word embeddings",
    "vector space",
    "cultural dimensions",
    "semantic relations",
    "class",
    "gender",
    "meaning",
    "corpus",
    "high-dimensional",
    "cultural associations",
    "neural-network"
   ],
   "phenomenon": "word embeddings (geometry of cultural meaning)",
   "claim_type": "method",
   "method": "analytic",
   "statistics": [
    {
     "term": "Bayesian",
     "evidence": "g models constructed with other algorithms, including Latent Semantic Analysis based on SVD (Dumais 2004) and Bayesian nonparametric approaches (Rudolph et al.",
     "decision": "accepted",
     "role": "substantive-mention",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/kozlowski2019-geometry-culture.html",
      "source_body_sha256": "6c09fd9ba83d93ec1e0c4d8ecc568430e08d8e933ffbe33b27c7733e20c1592d",
      "evidence_sha256": "07d1554ff9aade9e8c0619ccb5720c13b2f2453fdf5044df5074d72c9506b056",
      "grounding_status": "not-exactly-located"
     }
    },
    {
     "term": "bootstrap",
     "evidence": "Statistical Significance of Distances and Associations We propose well-established bootstrapping and subsampling methods to nonparametrically demonstrate the stability and signific",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/kozlowski2019-geometry-culture.html",
      "source_body_sha256": "6c09fd9ba83d93ec1e0c4d8ecc568430e08d8e933ffbe33b27c7733e20c1592d",
      "evidence_sha256": "82fba847c8f7d5580ab64fdf2cc4ff68e94bf98d0e58db894700a2f41068bac7",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "confidence-interval",
     "evidence": "For 90% confidence intervals, we randomly partition the corpus into 20 subcorpora, then calculate the error of our for each subsample s em",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/kozlowski2019-geometry-culture.html",
      "source_body_sha256": "6c09fd9ba83d93ec1e0c4d8ecc568430e08d8e933ffbe33b27c7733e20c1592d",
      "evidence_sha256": "85e1c92a59e626d89d789c3794711f2ab4e80b7df51264e469dae7bd803add3c",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "correlation",
     "evidence": "We test the correspondence between results from the survey and the word embedding model by calculating the Pearson’s correlation between a word’s mean rating on a given scale in th",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/kozlowski2019-geometry-culture.html",
      "source_body_sha256": "6c09fd9ba83d93ec1e0c4d8ecc568430e08d8e933ffbe33b27c7733e20c1592d",
      "evidence_sha256": "2ac42d93cbe50d2f35d1c074729a927d985723efaa438c879bd95e3f9bb5496c",
      "grounding_status": "not-exactly-located"
     }
    },
    {
     "term": "mean",
     "evidence": "imate the gender dimension of interest, we her ​ calculate a gender dimension by simply taking the arithmetic mean of a set of such pairs.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/kozlowski2019-geometry-culture.html",
      "source_body_sha256": "6c09fd9ba83d93ec1e0c4d8ecc568430e08d8e933ffbe33b27c7733e20c1592d",
      "evidence_sha256": "24fab3d0c83e89b8a2250d3600748fe839454673f1389bca5242bec74f23a63b",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "p-value",
     "evidence": "For all pairs 34 ​ ​ ​ ​ ​ of words that have a statistically significant difference in mean survey rating (p<0.01) for gender, class, and race associations within a substantive do",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit p-value or inferential p notation",
     "provenance": {
      "source_path": "papers/kozlowski2019-geometry-culture.html",
      "source_body_sha256": "6c09fd9ba83d93ec1e0c4d8ecc568430e08d8e933ffbe33b27c7733e20c1592d",
      "evidence_sha256": "81bc981e68e494356b5c6aba5eb700237778c1668a1cbe0cc8dfc8b0601e818d",
      "grounding_status": "not-exactly-located"
     }
    },
    {
     "term": "standard-deviation",
     "evidence": "n are (1) greater disagreement among those surveyed regarding race and class 33 ​ ​ ​ ​ associations (average standard deviation=23.1) than gender (SD=20.8); (2) greater disagreeme",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/kozlowski2019-geometry-culture.html",
      "source_body_sha256": "6c09fd9ba83d93ec1e0c4d8ecc568430e08d8e933ffbe33b27c7733e20c1592d",
      "evidence_sha256": "ac11895f5a48dc1556597e195b12417b7d7e484808ca2b8bcdc639f2bb130659",
      "grounding_status": "not-exactly-located"
     }
    },
    {
     "term": "statistical-significance",
     "evidence": "Moreover, turning text into data (Grimmer and Stewart 2013) requires credible methods for (1) evaluating the statistical significance of observed patterns and (2) disciplining the ",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit inferential-statistical context",
     "provenance": {
      "source_path": "papers/kozlowski2019-geometry-culture.html",
      "source_body_sha256": "6c09fd9ba83d93ec1e0c4d8ecc568430e08d8e933ffbe33b27c7733e20c1592d",
      "evidence_sha256": "610cdf6a651c871aa142a43bcb106a851026c2b9d9fd3020b878cd6d23effb20",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "variance",
     "evidence": "A key finding from this method is that much of the variance across all dimensions tested can be explained by just three core factors: Evaluation (good vs.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/kozlowski2019-geometry-culture.html",
      "source_body_sha256": "6c09fd9ba83d93ec1e0c4d8ecc568430e08d8e933ffbe33b27c7733e20c1592d",
      "evidence_sha256": "9ee32ffa70ba950562c1a77146e7ebb4587bb53d5eadbd578cfe245114253815",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "arXiv-nonexclusive-distrib-1.0",
   "license_url": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
   "source_url": "https://arxiv.org/abs/1803.09288",
   "doi": "10.1177/0003122419877135",
   "full_text_redistributable": false
  },
  {
   "id": "chernozhukov2018-double-ml",
   "title": "Double/Debiased Machine Learning for Treatment and Structural Parameters",
   "discipline": "statistics",
   "year": 2018,
   "authors": "Victor Chernozhukov, Denis Chetverikov, Mert Demirer, Esther Duflo",
   "url": "papers/chernozhukov2018-double-ml.html",
   "key_terms": [
    "nuisance",
    "parameters",
    "orthogonal",
    "cross-fitting",
    "treatment",
    "causal",
    "bias",
    "estimation",
    "machine learning"
   ],
   "phenomenon": "causal (treatment) effect estimation via machine learning (double/debiased ML)",
   "claim_type": "method",
   "method": "analytic",
   "statistics": [
    {
     "term": "confidence-interval",
     "evidence": "Then the confidence interval obeys (cid:104) CI := l′ ̃θ0 ± Φ−1(1 − α/2)(cid:112)l′ (cid:105) (cid:98)σ2l/N (cid:12) (cid:12) (cid:12)PP (l′θ0 ∈ CI) − (1 − α) (cid:12) (cid:12) (ci",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/chernozhukov2018-double-ml.html",
      "source_body_sha256": "14b5d220be48e046b401fc24d60723292f1bb366735458525d83432740369ed3",
      "evidence_sha256": "7114cd85edf87be2f7145f25630be935ed9d355b304193ec3ad0c8cb4ce40915",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "hypothesis-testing",
     "evidence": "Confidence intervals and hypothesis testing for high-dimensional regression.",
     "decision": "rejected",
     "role": "bibliography",
     "reason": "occurrence appears only in references",
     "provenance": {
      "source_path": "papers/chernozhukov2018-double-ml.html",
      "source_body_sha256": "14b5d220be48e046b401fc24d60723292f1bb366735458525d83432740369ed3",
      "evidence_sha256": "15b4a18a71e975cd626bcd6ac6a032b06dc494f8788640bf6d9585572651099b",
      "grounding_status": "verified-reference"
     }
    },
    {
     "term": "mean",
     "evidence": "We are now solving an auxiliary prediction problem to estimate the conditional mean of D given X, so we are doing “double prediction” or “double machine learning”.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "numeric mean or average in main text",
     "provenance": {
      "source_path": "papers/chernozhukov2018-double-ml.html",
      "source_body_sha256": "14b5d220be48e046b401fc24d60723292f1bb366735458525d83432740369ed3",
      "evidence_sha256": "9e23b3f580690a9fa68ddd7f5c8085e8c547eb588ab83f4d99c60495eb32fc49",
      "grounding_status": "not-exactly-located"
     }
    },
    {
     "term": "median",
     "evidence": "(Incorporating the Impact of Sample Splitting using Mean and Median Methods) For point estimation, we define ̃θmean 0 1 S S (cid:88) s=1 ̃θs 0 or ̃θmedian 0 = median{ ̃θs 0}S s=1, ",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/chernozhukov2018-double-ml.html",
      "source_body_sha256": "14b5d220be48e046b401fc24d60723292f1bb366735458525d83432740369ed3",
      "evidence_sha256": "adf01501b170de70163357dfeab9074024574dbca3c8ea9b1b1722294fd03310",
      "grounding_status": "not-exactly-located"
     }
    },
    {
     "term": "regression",
     "evidence": "ing it to provide theoretical properties of DML applied to learn the main regression parameter in a partially linear regression model, DML applied to learn the coefficient on an en",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/chernozhukov2018-double-ml.html",
      "source_body_sha256": "14b5d220be48e046b401fc24d60723292f1bb366735458525d83432740369ed3",
      "evidence_sha256": "0b940c25e6aa3a16f515a8393fdc56957bb1996e80acf0efa47d4992530c545d",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "sample-size",
     "evidence": "This bias results in the naive estimator failing to be N −1/2 consistent, where N is the sample size.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/chernozhukov2018-double-ml.html",
      "source_body_sha256": "14b5d220be48e046b401fc24d60723292f1bb366735458525d83432740369ed3",
      "evidence_sha256": "ffe805f1d2c7cccffb2b94a12258b2b6fa06571127e07435dd3616e13347c87a",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "standard-error",
     "evidence": "For comparison, we report two different standard errors.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit standard error or SEM",
     "provenance": {
      "source_path": "papers/chernozhukov2018-double-ml.html",
      "source_body_sha256": "14b5d220be48e046b401fc24d60723292f1bb366735458525d83432740369ed3",
      "evidence_sha256": "7ada15529039c6073fc2d11763224fc514e77ae552df1df1647cb724f5a72bb6",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "statistical-significance",
     "evidence": "Looking at the results, we see that the es- timated causal effect of 401(k) participation on net financial assets is uniformly positive and statistically significant across all of ",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit inferential-statistical context",
     "provenance": {
      "source_path": "papers/chernozhukov2018-double-ml.html",
      "source_body_sha256": "14b5d220be48e046b401fc24d60723292f1bb366735458525d83432740369ed3",
      "evidence_sha256": "8be90429a36d0faac86313d5e441742d1d6e53499f1e4c0aacfb659851c0df47",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "variance",
     "evidence": "The regularization in these estimators keeps the variance of the estimator from exploding but also necessarily induces substantive biases in the estimator (cid:98)g0 of g0.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/chernozhukov2018-double-ml.html",
      "source_body_sha256": "14b5d220be48e046b401fc24d60723292f1bb366735458525d83432740369ed3",
      "evidence_sha256": "cb7b7fdac258ea25908efa0234c2839e58f657746ab14c11b1fea947cb1e5fa0",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "arXiv-nonexclusive-distrib-1.0",
   "license_url": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
   "source_url": "https://arxiv.org/abs/1608.00060",
   "doi": "10.1111/ectj.12097",
   "full_text_redistributable": false
  },
  {
   "id": "eht2019-m87-black-hole",
   "title": "First M87 Event Horizon Telescope Results. I. The Shadow of the Supermassive Black Hole",
   "discipline": "astronomy",
   "year": 2019,
   "authors": "The Event Horizon Telescope Collaboration",
   "url": "papers/eht2019-m87-black-hole.html",
   "key_terms": [
    "black hole",
    "event horizon",
    "shadow",
    "supermassive",
    "accretion",
    "general relativity",
    "very long baseline interferometry",
    "photon ring",
    "emission ring",
    "relativistic beaming",
    "M87"
   ],
   "phenomenon": "black hole shadow (first event-horizon-scale image of M87*)",
   "claim_type": "existence",
   "method": "observation",
   "statistics": [
    {
     "term": "Bayesian",
     "evidence": "We used two distinct Bayesian-inference algorithms and demonstrate that such crescent models are statistically preferred over other comparably complex geometric models that we have",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/eht2019-m87-black-hole.html",
      "source_body_sha256": "df363ec94b768642fd21b9e9b5466924647237718f45a6aa5cf0c8efb49bb19a",
      "evidence_sha256": "71556395933653d949c7145b3f45719518f51542818a4ef6b08f344d24ee81bf",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "correlation",
     "evidence": "Data from the fringe-fitting pipelines were scaled from correlation coefficients to a uniform physical flux density scale (in Jansky) by using an independent apriori estimate of t",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/eht2019-m87-black-hole.html",
      "source_body_sha256": "df363ec94b768642fd21b9e9b5466924647237718f45a6aa5cf0c8efb49bb19a",
      "evidence_sha256": "b40d3497ced61ddfde9b233784390257b03cdec273055b358e073215b63c66f6",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "hypothesis-testing",
     "evidence": "our measurement of the black hole mass in M87* is not inconsistent with all of the prior mass measurements, this allows us to conclude that the null hypothesis of the Kerr metric (",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit hypothesis-testing language",
     "provenance": {
      "source_path": "papers/eht2019-m87-black-hole.html",
      "source_body_sha256": "df363ec94b768642fd21b9e9b5466924647237718f45a6aa5cf0c8efb49bb19a",
      "evidence_sha256": "5ebb7a8e68c566061ea85b45acf1479d3e2a812aa8339a739eca962cce591a34",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "mean",
     "evidence": "The image is the average of three different imaging methods after convolving each with a circular Gaussian kernel to give matched resolutions.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "numeric mean or average in main text",
     "provenance": {
      "source_path": "papers/eht2019-m87-black-hole.html",
      "source_body_sha256": "df363ec94b768642fd21b9e9b5466924647237718f45a6aa5cf0c8efb49bb19a",
      "evidence_sha256": "341da456afe22900bf0687d30554cbd72e2fed64d68eebade211e8581e0e92ea",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "median",
     "evidence": "median zenith atmospheric opacities at 230 GHz ranging from 0.03 to 0.28 over the different locations.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/eht2019-m87-black-hole.html",
      "source_body_sha256": "df363ec94b768642fd21b9e9b5466924647237718f45a6aa5cf0c8efb49bb19a",
      "evidence_sha256": "dec40033efdf67b26fde5e4c91ae55d1f6f5c16173f595187cf230c92a210206",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "null-hypothesis",
     "evidence": "our measurement of the black hole mass in M87* is not inconsistent with all of the prior mass measurements, this allows us to conclude that the null hypothesis of the Kerr metric (",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit null-hypothesis language",
     "provenance": {
      "source_path": "papers/eht2019-m87-black-hole.html",
      "source_body_sha256": "df363ec94b768642fd21b9e9b5466924647237718f45a6aa5cf0c8efb49bb19a",
      "evidence_sha256": "5ebb7a8e68c566061ea85b45acf1479d3e2a812aa8339a739eca962cce591a34",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "variance",
     "evidence": "e, except for the a*=−0.94 MAD models, which fail to produce images that are sufficiently stable (i.e., the variance among snapshots is too large to be statistically consistent w",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/eht2019-m87-black-hole.html",
      "source_body_sha256": "df363ec94b768642fd21b9e9b5466924647237718f45a6aa5cf0c8efb49bb19a",
      "evidence_sha256": "617c804f9c977e3643eb7b62ff2d93f364d13aa25d17498880a888beeff17d30",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "CC-BY-3.0",
   "license_url": "http://creativecommons.org/licenses/by/3.0/",
   "source_url": "https://arxiv.org/abs/1906.11238",
   "doi": "10.3847/2041-8213/ab0ec7",
   "full_text_redistributable": true
  },
  {
   "id": "preskill2018-nisq",
   "title": "Quantum Computing in the NISQ Era and Beyond",
   "discipline": "quantum-computing",
   "year": 2018,
   "authors": "John Preskill",
   "url": "papers/preskill2018-nisq.html",
   "key_terms": [
    "quantum",
    "qubits",
    "NISQ",
    "noise",
    "classical",
    "fault-tolerant",
    "computers",
    "entanglement",
    "supremacy"
   ],
   "phenomenon": "noisy intermediate-scale quantum (NISQ) computing (near-term, noise-limited quantum devices)",
   "claim_type": "existence",
   "method": "review",
   "statistics": [
    {
     "term": "correlation",
     "evidence": "Laughlin and Pines [8] emphasized in the statement I quoted earlier, we think that simulating such “strongly correlated” matter is a difficult computational problem, because many v",
     "decision": "rejected",
     "role": "non-statistical-sense",
     "reason": "lexical match does not denote the controlled statistical concept",
     "provenance": {
      "source_path": "papers/preskill2018-nisq.html",
      "source_body_sha256": "e5357fcbafb42b4b2bd6a46cec21fc67cfd81dd140ad7cffd7a9d4d0e710770c",
      "evidence_sha256": "a29208e487042d3880ef5153796ac2aa4a8773810eda6422e6c593022e67bb89",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "CC-BY-4.0",
   "license_url": "http://creativecommons.org/licenses/by/4.0/",
   "source_url": "https://arxiv.org/abs/1801.00862",
   "doi": "10.22331/q-2018-08-06-79",
   "full_text_redistributable": true
  },
  {
   "id": "geim2007-rise-graphene",
   "title": "The Rise of Graphene",
   "discipline": "materials-science",
   "year": 2007,
   "authors": "A. K. Geim, K. S. Novoselov",
   "url": "papers/geim2007-rise-graphene.html",
   "key_terms": [],
   "phenomenon": "",
   "claim_type": "",
   "method": "review",
   "statistics": [
    {
     "term": "mean",
     "evidence": "Bilayer graphene: short range intera tions in 1956) sin e mean (cid:28)eld al ulations give a riti al value for U to be of the order of U/t 1.7.",
     "decision": "rejected",
     "role": "non-statistical-sense",
     "reason": "lexical match does not denote the controlled statistical concept",
     "provenance": {
      "source_path": "papers/geim2007-rise-graphene.html",
      "source_body_sha256": null,
      "evidence_sha256": "ad7dc913c58c5cf76d4b7c274549ab6188a3b48ac3e6aa96777f8cbd8f13ccfc",
      "grounding_status": "source-unavailable"
     }
    },
    {
     "term": "null-hypothesis",
     "evidence": "Hi = V Ni j=1 X δ(r rj)I (192) The formulation of the problem in se ond quantization requires the solution of H0 , whi h was done in Se tion II.I.",
     "decision": "rejected",
     "role": "symbol-collision",
     "reason": "H0 denotes a Hubble parameter, Hamiltonian, operator, or generating function rather than a null hypothesis",
     "provenance": {
      "source_path": "papers/geim2007-rise-graphene.html",
      "source_body_sha256": null,
      "evidence_sha256": "66230c26702920d23f77ba8341b0c3a7b8ab67770bed465e551466c2d9a67b4b",
      "grounding_status": "source-unavailable"
     }
    }
   ],
   "license_spdx": "arxiv-open",
   "source_url": "https://arxiv.org/abs/0709.1163",
   "doi": "10.1038/nmat1849",
   "full_text_redistributable": false
  },
  {
   "id": "grover2016-node2vec",
   "title": "node2vec: Scalable Feature Learning for Networks",
   "discipline": "network-science",
   "year": 2016,
   "authors": "Aditya Grover, Jure Leskovec",
   "url": "papers/grover2016-node2vec.html",
   "key_terms": [
    "node2vec",
    "network",
    "feature",
    "embedding",
    "neighborhood",
    "random walk",
    "nodes",
    "representation",
    "biased"
   ],
   "phenomenon": "word embeddings (feature learning for graph nodes)",
   "claim_type": "method",
   "method": "experiment",
   "statistics": [
    {
     "term": "p-value",
     "evidence": "at our experiments for 10 random seed initial- izations, and our results are statistically significant with a p-value of less than 0.01.The best in-out and return hyperparameters w",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit p-value or inferential p notation",
     "provenance": {
      "source_path": "papers/grover2016-node2vec.html",
      "source_body_sha256": "caa5f542d203826b3798e51ec30379954f9652bba3610666a2f520556838ef65",
      "evidence_sha256": "5b45f67ff599ea2570e68351e44a95dc214e4d40e257b832805aa9c342a48e6d",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "regression",
     "evidence": "The node feature representations are input to a one-vs-rest logistic regression classifier with L2 regularization.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/grover2016-node2vec.html",
      "source_body_sha256": "caa5f542d203826b3798e51ec30379954f9652bba3610666a2f520556838ef65",
      "evidence_sha256": "f2d3b3a4e0327668dacd65e68d90524f90ff2f3efabd358e4269add85c65c57f",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "sample-size",
     "evidence": "The opposite is true for DFS which can explore larger parts of the network as it can move further away from the source node u (with sample size k being fixed).",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/grover2016-node2vec.html",
      "source_body_sha256": "caa5f542d203826b3798e51ec30379954f9652bba3610666a2f520556838ef65",
      "evidence_sha256": "7d44a285867e6dda5987b83e6be78598f524a16a41c4bb577c3237234e31a50a",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "statistical-significance",
     "evidence": "We repeat our experiments for 10 random seed initial- izations, and our results are statistically significant with a p-value of less than 0.01.The best in-out and return hyperparam",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit inferential-statistical context",
     "provenance": {
      "source_path": "papers/grover2016-node2vec.html",
      "source_body_sha256": "caa5f542d203826b3798e51ec30379954f9652bba3610666a2f520556838ef65",
      "evidence_sha256": "12a636a62b1362ee3e026cdb79f891ad9d4fbd77d567d64d0afaa113a708135b",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "arXiv-nonexclusive-distrib-1.0",
   "license_url": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
   "source_url": "https://arxiv.org/abs/1607.00653",
   "doi": "10.1145/2939672.2939754",
   "full_text_redistributable": false
  },
  {
   "id": "asgari2015-protvec",
   "title": "Continuous Distributed Representation of Biological Sequences for Deep Proteomics and Genomics",
   "discipline": "biology",
   "year": 2015,
   "authors": "Ehsaneddin Asgari, Mohammad R. K. Mofrad",
   "url": "papers/asgari2015-protvec.html",
   "key_terms": [
    "protein",
    "sequence",
    "vector",
    "n-gram",
    "embedding",
    "representation",
    "classification",
    "word2vec"
   ],
   "phenomenon": "word embeddings (distributed representations of biological sequences; ProtVec)",
   "claim_type": "method",
   "method": "experiment",
   "statistics": [
    {
     "term": "regression",
     "evidence": "Least squares support vector machine classifiers.",
     "decision": "rejected",
     "role": "non-statistical-sense",
     "reason": "lexical match does not denote the controlled statistical concept",
     "provenance": {
      "source_path": "papers/asgari2015-protvec.html",
      "source_body_sha256": "33cee34a443aa398762d5c60199784b71cd6f5095cb02ca2075489983bf5aed8",
      "evidence_sha256": "e22dea144a209ad0495b6ab2ea5bdf5f77605aef395b047c69cb53332fe5bc63",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "statistical-significance",
     "evidence": "A visual comparison between Figs 3 and 2 suggest that the FG-Nups have a significantly higher amount of hydrophobic residues and less polar residues in their PLOS ONE | DOI:10.1371",
     "decision": "rejected",
     "role": "lexical-significance",
     "reason": "significant/significantly is used without inferential-statistical meaning",
     "provenance": {
      "source_path": "papers/asgari2015-protvec.html",
      "source_body_sha256": "33cee34a443aa398762d5c60199784b71cd6f5095cb02ca2075489983bf5aed8",
      "evidence_sha256": "63d1655e12b5fd552ff8ac2b926451b74af119a7373a1e995ad32601fc166d14",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "CC-BY-4.0",
   "license_url": "http://creativecommons.org/licenses/by/4.0/",
   "source_url": "https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0141287",
   "doi": "10.1371/journal.pone.0141287",
   "full_text_redistributable": true
  },
  {
   "id": "gilmer2017-message-passing",
   "title": "Neural Message Passing for Quantum Chemistry",
   "discipline": "chemistry",
   "year": 2017,
   "authors": "Justin Gilmer, Samuel S. Schoenholz, Patrick F. Riley, Oriol Vinyals, George E. Dahl",
   "url": "papers/gilmer2017-message-passing.html",
   "key_terms": [
    "message passing",
    "quantum chemistry",
    "molecular property",
    "graph isomorphism",
    "MPNN",
    "DFT",
    "chemical accuracy",
    "readout function",
    "node features",
    "hidden states",
    "supervised learning"
   ],
   "phenomenon": "molecular property prediction with graph neural networks (message passing)",
   "claim_type": "method",
   "method": "experiment",
   "statistics": [
    {
     "term": "mean",
     "evidence": "Note that the Lvw are all scalar valued, so this model corre- sponds to taking a certain weighted average of neighboring nodes at each time step.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "numeric mean or average in main text",
     "provenance": {
      "source_path": "papers/gilmer2017-message-passing.html",
      "source_body_sha256": "da5b9bf13c24e3b4bd6fcb39d8ada73eb2e16ba493d323be53088c9f436d8092",
      "evidence_sha256": "5462cad57ff145e9cdb448b1e0a367c7f250bf8671de0ed4174327930aedabb4",
      "grounding_status": "verified-reference"
     }
    },
    {
     "term": "sample-size",
     "evidence": "the edge network + set2set model on different sized training sets (N denotes the number of training samples) N=11k N=35k N=58k N=82k N=110k Target 1.28 mu 2.76 alpha HOMO 2.33 LUMO",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/gilmer2017-message-passing.html",
      "source_body_sha256": "da5b9bf13c24e3b4bd6fcb39d8ada73eb2e16ba493d323be53088c9f436d8092",
      "evidence_sha256": "44fb52cdce6c8e64aa6e1a525bb14f1fbd023ea1ada3bc83a9d7f7caa936e4d6",
      "grounding_status": "verified-reference"
     }
    }
   ],
   "license_spdx": "arXiv-nonexclusive-distrib-1.0",
   "license_url": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
   "source_url": "https://arxiv.org/abs/1704.01212",
   "doi": "",
   "full_text_redistributable": false
  },
  {
   "id": "candes2009-matrix-completion",
   "title": "Exact Matrix Completion via Convex Optimization",
   "discipline": "mathematics",
   "year": 2009,
   "authors": "Emmanuel J. Candès, Benjamin Recht",
   "url": "papers/candes2009-matrix-completion.html",
   "key_terms": [
    "matrix",
    "completion",
    "low-rank",
    "nuclear norm",
    "convex",
    "recovery",
    "entries",
    "incoherence",
    "rank"
   ],
   "phenomenon": "matrix completion (exact low-rank recovery from partial entries; the Netflix problem)",
   "claim_type": "existence",
   "method": "analytic",
   "statistics": [
    {
     "term": "chi-square",
     "evidence": "Lemma 2.3 Let Yd be distributed as a chi-squared random variable with d degrees of freedom.",
     "decision": "accepted",
     "role": "substantive-mention",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/candes2009-matrix-completion.html",
      "source_body_sha256": "e84e92ea49c3c166c7f03174a6f197a008c3cd3c224c9d481aab3d9c6a548a25",
      "evidence_sha256": "b39518c9f32f2a24ca650165993ac60e25fc9919aa26c455080819a38f0226ce",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "degrees-of-freedom",
     "evidence": "Lemma 2.3 Let Yd be distributed as a chi-squared random variable with d degrees of freedom.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "degrees of freedom in a statistical model/test",
     "provenance": {
      "source_path": "papers/candes2009-matrix-completion.html",
      "source_body_sha256": "e84e92ea49c3c166c7f03174a6f197a008c3cd3c224c9d481aab3d9c6a548a25",
      "evidence_sha256": "b39518c9f32f2a24ca650165993ac60e25fc9919aa26c455080819a38f0226ce",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "sample-size",
     "evidence": "matrices from on the order of nr—up to logarithmic multiplicative factors—randomly selected entries? Can the sample size be merely proportional to the true complexity of the low-ra",
     "decision": "accepted",
     "role": "substantive-mention",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/candes2009-matrix-completion.html",
      "source_body_sha256": "e84e92ea49c3c166c7f03174a6f197a008c3cd3c224c9d481aab3d9c6a548a25",
      "evidence_sha256": "40a400abfef474309093bdd10e3ed469489342309c7b98600231a9b3ae36d94a",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "statistical-significance",
     "evidence": "A common hypothesis in the multitask setting is that the wk corresponding to each of the classes together span a very low dimensional subspace with dimension significantly smaller ",
     "decision": "rejected",
     "role": "lexical-significance",
     "reason": "significant/significantly is used without inferential-statistical meaning",
     "provenance": {
      "source_path": "papers/candes2009-matrix-completion.html",
      "source_body_sha256": "e84e92ea49c3c166c7f03174a6f197a008c3cd3c224c9d481aab3d9c6a548a25",
      "evidence_sha256": "f7a591bdaaf051d5b079a0db1772de9daddca9265d4eebc2ec086c654c1b7674",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "variance",
     "evidence": "|Yk| ≤ c, (cid:32) n (cid:88) k=1 (cid:33) Yk| > t ≤ 2e−t2/(2σ2+2ct/3), (6.16) 28 where σ2 is the sum of the variances, σ2 ≡ (cid:80)n k=1 Var(Yk).",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/candes2009-matrix-completion.html",
      "source_body_sha256": "e84e92ea49c3c166c7f03174a6f197a008c3cd3c224c9d481aab3d9c6a548a25",
      "evidence_sha256": "63b388c3a3389694dfc4753e8ec5d800b8851094d05a2241d1e3a0b2690da1b4",
      "grounding_status": "not-exactly-located"
     }
    }
   ],
   "license_spdx": "arXiv-nonexclusive-distrib-1.0",
   "license_url": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
   "source_url": "https://arxiv.org/abs/0805.4471",
   "doi": "10.1007/s10208-009-9045-5",
   "full_text_redistributable": false
  },
  {
   "id": "perlmutter1999-omega-lambda",
   "title": "Measurements of Omega and Lambda from 42 High-Redshift Supernovae",
   "discipline": "astronomy",
   "year": 1999,
   "authors": "Saul Perlmutter, The Supernova Cosmology Project",
   "url": "papers/perlmutter1999-omega-lambda.html",
   "key_terms": [
    "supernovae",
    "redshift",
    "cosmological",
    "Omega",
    "Lambda",
    "universe",
    "expansion",
    "accelerating",
    "density"
   ],
   "phenomenon": "accelerating expansion of the universe (dark energy / cosmological constant)",
   "claim_type": "existence",
   "method": "observation",
   "statistics": [
    {
     "term": "Bayesian",
     "evidence": "We have compared the results of Bayesian and classical, “fre- quentist,” fitting procedures.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/perlmutter1999-omega-lambda.html",
      "source_body_sha256": "42aafa766f9c38e2295de101df39d01291c4db1d389a9caf1f4fa46575e6e19c",
      "evidence_sha256": "1ff1a36552310176f706cfb82e41ec8749000f8ec58e364019b5bc2a62e1b645",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "chi-square",
     "evidence": "We have therefore performed the fits two ways: minimizing χ2 measured using either mag- nitude residuals or flux residuals.",
     "decision": "accepted",
     "role": "substantive-mention",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/perlmutter1999-omega-lambda.html",
      "source_body_sha256": "42aafa766f9c38e2295de101df39d01291c4db1d389a9caf1f4fa46575e6e19c",
      "evidence_sha256": "4203bc9bbf1ef3beb3a93a8d21c41b1dfff577f869724cf5db62c32fdef1eeb0",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "correlation",
     "evidence": "however, are typically detected be- fore maximum, and their detection magnitudes and peak mag- nitudes have a correlation coefficient of only 0.35, so the ef- fects of classical Ma",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/perlmutter1999-omega-lambda.html",
      "source_body_sha256": "42aafa766f9c38e2295de101df39d01291c4db1d389a9caf1f4fa46575e6e19c",
      "evidence_sha256": "2ca331e6561e420f795ad5f26a4d61379c2236135b8087889f0cba160ab69873",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "degrees-of-freedom",
     "evidence": "The main difference be- tween the fits is the goodness-of-fit: the larger χ2 per degree of freedom for Fit A, χ2 ν = 1.76, indicates that the outlier super- novae included in this ",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "degrees of freedom in a statistical model/test",
     "provenance": {
      "source_path": "papers/perlmutter1999-omega-lambda.html",
      "source_body_sha256": "42aafa766f9c38e2295de101df39d01291c4db1d389a9caf1f4fa46575e6e19c",
      "evidence_sha256": "aa192268696a979582faf39d8319b8df9933d7def7f1f2de292f0cedefc93e79",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "mean",
     "evidence": "it removes the faintest of the high- redshift supernovae, but it does ensure that the error-weighted E(B- V) mean of the remaining supernova subset is a good indicator of any redde",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "numeric mean or average in main text",
     "provenance": {
      "source_path": "papers/perlmutter1999-omega-lambda.html",
      "source_body_sha256": "42aafa766f9c38e2295de101df39d01291c4db1d389a9caf1f4fa46575e6e19c",
      "evidence_sha256": "b07ada2f61a606909eab6283a8ff918483fad8d6dc803f6f39b600809f446274",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "median",
     "evidence": "17%, for any high-redshift value of B R median SCP i E(B- V ) h Note that the modes of both distributions appear to be at zero reddening, and similarly the",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/perlmutter1999-omega-lambda.html",
      "source_body_sha256": "42aafa766f9c38e2295de101df39d01291c4db1d389a9caf1f4fa46575e6e19c",
      "evidence_sha256": "ca5beb2ffb64a8aa9b1ec3874c09ea022f9a22fc31493b4765ab3307e87526d8",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "null-hypothesis",
     "evidence": "For any value of the Hubble constant less than H0 = 70 km s- 1 Mpc- 1, the implied age of the universe is greater than 13 Gyr, allowing enough time for the oldest stars in globular",
     "decision": "rejected",
     "role": "symbol-collision",
     "reason": "H0 denotes a Hubble parameter, Hamiltonian, operator, or generating function rather than a null hypothesis",
     "provenance": {
      "source_path": "papers/perlmutter1999-omega-lambda.html",
      "source_body_sha256": "42aafa766f9c38e2295de101df39d01291c4db1d389a9caf1f4fa46575e6e19c",
      "evidence_sha256": "e01e1f4c0195aeaf53ec3c98f362de70670479214b2ccb1a33716d5bd2e59f31",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "sample-size",
     "evidence": "Section 2 are not a significant factor in arriving at this agree- ment.) With the small P97 sample size of seven supernovae (only five of which were used in the P97 width-corrected",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/perlmutter1999-omega-lambda.html",
      "source_body_sha256": "42aafa766f9c38e2295de101df39d01291c4db1d389a9caf1f4fa46575e6e19c",
      "evidence_sha256": "4c9ab7ee1a967444917c0f1b52fa9c1c8ef605d0870398a66d5386afcf5db175",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "standard-deviation",
     "evidence": ".) The confidence regions of Figure 5(c) and the Ωflat M results in Table 3 show that the cosmological parameters found for Fit H differ by less than half of a standard deviation f",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/perlmutter1999-omega-lambda.html",
      "source_body_sha256": "42aafa766f9c38e2295de101df39d01291c4db1d389a9caf1f4fa46575e6e19c",
      "evidence_sha256": "b698648720ca39435fbddabcfbc2063e086c0643ccf0f9c4105a052084cce456",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "statistical-significance",
     "evidence": "The high-redshift sample is now large enough that these subsets each contain enough supernovae to yield results of high statistical significance.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit inferential-statistical context",
     "provenance": {
      "source_path": "papers/perlmutter1999-omega-lambda.html",
      "source_body_sha256": "42aafa766f9c38e2295de101df39d01291c4db1d389a9caf1f4fa46575e6e19c",
      "evidence_sha256": "bd2ca4945c7b8730797c28e9374bfc8ad91948886c5588ccce6592bbc3a3a37a",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "arxiv-open",
   "license_url": "",
   "source_url": "https://arxiv.org/abs/astro-ph/9812133",
   "doi": "10.1086/307221",
   "full_text_redistributable": false
  },
  {
   "id": "recht2019-continuous-control",
   "title": "A Tour of Reinforcement Learning: The View from Continuous Control",
   "discipline": "control-theory",
   "year": 2019,
   "authors": "Benjamin Recht",
   "url": "papers/recht2019-continuous-control.html",
   "key_terms": [
    "reinforcement learning",
    "continuous control",
    "optimal control",
    "dynamical system",
    "policy",
    "linear quadratic regulator",
    "model-free",
    "model-based",
    "trajectory",
    "reward",
    "control action"
   ],
   "phenomenon": "continuous control (bridging reinforcement learning and optimal control)",
   "claim_type": "method",
   "method": "review",
   "statistics": [
    {
     "term": "Bayesian",
     "evidence": "Bayesian Optimal Control of Smoothly Parameterized Systems: The Lazy Posterior Sampling Algorithm.",
     "decision": "rejected",
     "role": "bibliography",
     "reason": "occurrence is confined to a citation/reference rather than the article body",
     "provenance": {
      "source_path": "papers/recht2019-continuous-control.html",
      "source_body_sha256": "e95e4c7ea1c1911e710bf0924678c12256492aa6d8bf6ec5b514c85f2839dc5e",
      "evidence_sha256": "5a6f03548f42086440b154aad967f41be36d9eaca54fae9ce6ad954091a077db",
      "grounding_status": "verified-reference"
     }
    },
    {
     "term": "bootstrap",
     "evidence": "The green curve depicts what happens when we estimate this difference between the models using a bootstrap simulation [29, 67].",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/recht2019-continuous-control.html",
      "source_body_sha256": "e95e4c7ea1c1911e710bf0924678c12256492aa6d8bf6ec5b514c85f2839dc5e",
      "evidence_sha256": "0e68562c1e45e3b5a3a048c5302c93b1acd2bac7da5ab4796be272e9aa6959a1",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "regression",
     "evidence": "This form of machine learning includes classification and regression as special cases.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/recht2019-continuous-control.html",
      "source_body_sha256": "e95e4c7ea1c1911e710bf0924678c12256492aa6d8bf6ec5b514c85f2839dc5e",
      "evidence_sha256": "36714fb5364d3bfa6ea616cc9f59096a64ee2db07472b1e071f3b4f0a6fa82e9",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "variance",
     "evidence": "One can reduce the variance of such a finite-difference estimate by sampling along multiple random directions and averaging: g(m) σ (θ) = 1 m m (cid:88) i=1 R(θ + σ(cid:15)i) − R(θ",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/recht2019-continuous-control.html",
      "source_body_sha256": "e95e4c7ea1c1911e710bf0924678c12256492aa6d8bf6ec5b514c85f2839dc5e",
      "evidence_sha256": "13b335e16c82fd667bdee365f7a1d9ea76f56c9c99e2e6f02fe62f6253ca18cf",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "arXiv-nonexclusive-distrib-1.0",
   "license_url": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
   "source_url": "https://arxiv.org/abs/1806.09460",
   "doi": "10.1146/annurev-control-053018-023825",
   "full_text_redistributable": false
  },
  {
   "id": "bengio2021-ml-combinatorial",
   "title": "Machine Learning for Combinatorial Optimization: A Methodological Tour d'Horizon",
   "discipline": "operations-research",
   "year": 2021,
   "authors": "Yoshua Bengio, Andrea Lodi, Antoine Prouvost",
   "url": "papers/bengio2021-ml-combinatorial.html",
   "key_terms": [
    "combinatorial optimization",
    "machine learning",
    "operations research",
    "traveling salesman problem",
    "decision variables",
    "NP-hard",
    "deep learning",
    "reinforcement learning",
    "heuristics",
    "generalizing",
    "policy"
   ],
   "phenomenon": "machine learning for combinatorial optimization",
   "claim_type": "existence",
   "method": "review",
   "statistics": [],
   "license_spdx": "CC-BY-NC-SA-4.0",
   "license_url": "http://creativecommons.org/licenses/by-nc-sa/4.0/",
   "source_url": "https://arxiv.org/abs/1811.06128",
   "doi": "10.1016/j.ejor.2020.07.063",
   "full_text_redistributable": false
  },
  {
   "id": "karras2019-stylegan",
   "title": "A Style-Based Generator Architecture for Generative Adversarial Networks",
   "discipline": "art-history",
   "year": 2019,
   "authors": "Tero Karras, Samuli Laine, Timo Aila",
   "url": "papers/karras2019-stylegan.html",
   "key_terms": [
    "generator",
    "latent",
    "style",
    "disentangle",
    "images",
    "synthesis",
    "interpolation",
    "network",
    "AdaIN"
   ],
   "phenomenon": "generative image synthesis (style-based GAN generator)",
   "claim_type": "method",
   "method": "experiment",
   "statistics": [
    {
     "term": "Bayesian",
     "evidence": "Auto-encoding variational bayes.",
     "decision": "rejected",
     "role": "bibliography",
     "reason": "occurrence is confined to a citation/reference rather than the article body",
     "provenance": {
      "source_path": "papers/karras2019-stylegan.html",
      "source_body_sha256": "3e42af42576e52f7f88466ad55796b4da1f7c84b7b1eb8a28be998eda6bc4e35",
      "evidence_sha256": "80f5384c0c88c6195937a8f529b300b4c65e554b39c34a714ba6e15e57a49c60",
      "grounding_status": "verified-reference"
     }
    },
    {
     "term": "mean",
     "evidence": "In particu- lar, we use the same discriminator architecture, resolution- dependent minibatch sizes, Adam [33] hyperparameters, and exponential moving average of the generator.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/karras2019-stylegan.html",
      "source_body_sha256": "3e42af42576e52f7f88466ad55796b4da1f7c84b7b1eb8a28be998eda6bc4e35",
      "evidence_sha256": "864916695107ffeeebceff6b633044bbc9cadd68487960e08cf70926e8d7bb6f",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "standard-deviation",
     "evidence": "(a) Generated image (b) Stochastic variation (c) Standard deviation (a) Two generated Figure 4.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/karras2019-stylegan.html",
      "source_body_sha256": "3e42af42576e52f7f88466ad55796b4da1f7c84b7b1eb8a28be998eda6bc4e35",
      "evidence_sha256": "a61f83b10d6293e4c7feda7ec8ff9f87e8a3e3bd75048ec385a3c302570d603b",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "arXiv-nonexclusive-distrib-1.0",
   "license_url": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
   "source_url": "https://arxiv.org/abs/1812.04948",
   "doi": "",
   "full_text_redistributable": false
  },
  {
   "id": "ugander2011-facebook-graph",
   "title": "The Anatomy of the Facebook Social Graph",
   "discipline": "sociology",
   "year": 2011,
   "authors": "Johan Ugander, Brian Karrer, Lars Backstrom, Cameron Marlow",
   "url": "papers/ugander2011-facebook-graph.html",
   "key_terms": [
    "graph",
    "users",
    "network",
    "clustering",
    "degree",
    "friendships",
    "assortativity",
    "neighborhoods",
    "social"
   ],
   "phenomenon": "social network structure (the Facebook graph; small-world / six degrees of separation)",
   "claim_type": "existence",
   "method": "observation",
   "statistics": [
    {
     "term": "correlation",
     "evidence": "For the Facebook network, r = 0.226, displaying positive correlations with similar magnitude to other social graphs.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/ugander2011-facebook-graph.html",
      "source_body_sha256": "9e139febd2b9a8b17659c52d15484c1667d91a8cfafc84a2cbd5337cf5bb77ca",
      "evidence_sha256": "4fb4354365dce715d201f661cad26d4611a2960ada8b9a706f891948e4806f47",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "mean",
     "evidence": "6a, the mean of these distributions is clearly less than the mean of the orange distribution which represents fol- lowing a random edge in the network, except for the green line de",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/ugander2011-facebook-graph.html",
      "source_body_sha256": "9e139febd2b9a8b17659c52d15484c1667d91a8cfafc84a2cbd5337cf5bb77ca",
      "evidence_sha256": "a70bc1f00a267acddacad6dc6a99ac4006ca0baea7f9db68d135e2162c3aaa70",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "median",
     "evidence": "For example, for users with 100 friends, the average local clustering coefficient is 0.14, indicating that for a median user, 14% of all their friend pairs are them- selves friends",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/ugander2011-facebook-graph.html",
      "source_body_sha256": "9e139febd2b9a8b17659c52d15484c1667d91a8cfafc84a2cbd5337cf5bb77ca",
      "evidence_sha256": "5e811281c3b7a7b1e11c3b007c70f29b16f86f3539754423b2c21e68c23d09a7",
      "grounding_status": "not-exactly-located"
     }
    },
    {
     "term": "sample-size",
     "evidence": "According to our definition of active, the population of active Facebook users consisted of around n = 721 million individuals at the time of our measurements.",
     "decision": "rejected",
     "role": "non-statistical-sense",
     "reason": "lexical match does not denote the controlled statistical concept",
     "provenance": {
      "source_path": "papers/ugander2011-facebook-graph.html",
      "source_body_sha256": "9e139febd2b9a8b17659c52d15484c1667d91a8cfafc84a2cbd5337cf5bb77ca",
      "evidence_sha256": "7037ad1f2e41cfbf628c3a707aa69efd463e5831668953e0bc325ce365ff619a",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "statistical-significance",
     "evidence": "Restricting the comparison to individuals age 13 or more and with access to the internet (the set of individuals eligible to have Facebook accounts) would put this percentage signi",
     "decision": "rejected",
     "role": "lexical-significance",
     "reason": "significant/significantly is used without inferential-statistical meaning",
     "provenance": {
      "source_path": "papers/ugander2011-facebook-graph.html",
      "source_body_sha256": "9e139febd2b9a8b17659c52d15484c1667d91a8cfafc84a2cbd5337cf5bb77ca",
      "evidence_sha256": "83eaf46871ee3c46215f4361bd0baa96a32e3a1a2e9c00aa5a54ecfe897f2f8b",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "arXiv-nonexclusive-distrib-1.0",
   "license_url": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
   "source_url": "https://arxiv.org/abs/1111.4503",
   "doi": "",
   "full_text_redistributable": false
  },
  {
   "id": "gerardvaret2010-prandtl-illposedness",
   "title": "On the Ill-Posedness of the Prandtl Equation",
   "discipline": "mathematics",
   "year": 2010,
   "authors": "David Gérard-Varet, Emmanuel Dormy",
   "url": "papers/gerardvaret2010-prandtl-illposedness.html",
   "key_terms": [
    "Prandtl",
    "boundary layer",
    "ill-posed",
    "instability",
    "Sobolev",
    "viscous",
    "equation",
    "Navier-Stokes",
    "analytic"
   ],
   "phenomenon": "well-posedness of the Prandtl boundary-layer equation",
   "claim_type": "existence",
   "method": "analytic",
   "statistics": [],
   "license_spdx": "arXiv-nonexclusive-distrib-1.0",
   "license_url": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
   "source_url": "https://arxiv.org/abs/0904.0434",
   "doi": "10.1090/S0894-0347-09-00652-3",
   "full_text_redistributable": false
  },
  {
   "id": "clauset2009-power-law-data",
   "title": "Power-Law Distributions in Empirical Data",
   "discipline": "statistics",
   "year": 2009,
   "authors": "Aaron Clauset, Cosma Rohilla Shalizi, M. E. J. Newman",
   "url": "papers/clauset2009-power-law-data.html",
   "key_terms": [
    "power law",
    "power-law distribution",
    "maximum likelihood",
    "scaling parameter",
    "exponent",
    "goodness-of-fit",
    "Kolmogorov-Smirnov",
    "heavy-tailed",
    "tail",
    "empirical data"
   ],
   "phenomenon": "power-law distributions (heavy-tailed distributions)",
   "claim_type": "method",
   "method": "analytic",
   "statistics": [
    {
     "term": "Bayesian",
     "evidence": "ny of several other established and statistically principled approaches for model comparison, such as a fully Bayesian approach [32], a cross-validation approach [59], or a minimum",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/clauset2009-power-law-data.html",
      "source_body_sha256": "fdd55be2ef634e71c4c4773765c89de3575b87479578140c6812ce0542d9e36c",
      "evidence_sha256": "5dd8c3740ef06fe9fa9240c66b5cc8776ad471ff77cd7a2ca0d474bbae474ef7",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "confidence-interval",
     "evidence": "ack line shows the rate of misclassification (over 1000 repetitions) of power law samples as log-normals (95% confidence interval shown in grey), while the blue (dashed) line shows",
     "decision": "accepted",
     "role": "substantive-mention",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/clauset2009-power-law-data.html",
      "source_body_sha256": "fdd55be2ef634e71c4c4773765c89de3575b87479578140c6812ce0542d9e36c",
      "evidence_sha256": "f124fecd4b6b8728bd4e5e4c9c3a28c40faa12d14c4efbf9e19a12a7890602b0",
      "grounding_status": "not-exactly-located"
     }
    },
    {
     "term": "correlation",
     "evidence": "Furthermore, the assumption of independence now fails, because P (x) = P (x + 1) + p(x) and hence adjacent values of the CDF are strongly correlated.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/clauset2009-power-law-data.html",
      "source_body_sha256": "fdd55be2ef634e71c4c4773765c89de3575b87479578140c6812ce0542d9e36c",
      "evidence_sha256": "cc3ef07bff18938f58d5ab5bce1c1c6533cb4881075ff027526756cd41aba726",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "hypothesis-testing",
     "evidence": "In the terminology of statistical theory, the value of r2 has very little power as a hypothesis test because the probability of successfully detecting a violation of the power-law ",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit hypothesis-testing language",
     "provenance": {
      "source_path": "papers/clauset2009-power-law-data.html",
      "source_body_sha256": "fdd55be2ef634e71c4c4773765c89de3575b87479578140c6812ce0542d9e36c",
      "evidence_sha256": "0cd78dd34e57b4fa5c05ab82a937aa278c717df9200c10f100e516b348c95be3",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "mean",
     "evidence": "Mean of the maximum likelihood estimate for the scaling parameter for 5000 samples drawn from the test distribution, Eq.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/clauset2009-power-law-data.html",
      "source_body_sha256": "fdd55be2ef634e71c4c4773765c89de3575b87479578140c6812ce0542d9e36c",
      "evidence_sha256": "c87f40c456cb2e60b7df9e12c07358d229e7f06413902732244512b0a4fbac5e",
      "grounding_status": "not-exactly-located"
     }
    },
    {
     "term": "null-hypothesis",
     "evidence": "Normally one then considers low values of p to be good, since they indicate that the null hypothesis is unlikely to be correct.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit null-hypothesis language",
     "provenance": {
      "source_path": "papers/clauset2009-power-law-data.html",
      "source_body_sha256": "fdd55be2ef634e71c4c4773765c89de3575b87479578140c6812ce0542d9e36c",
      "evidence_sha256": "426d1b6b4da19944d06026d760b50b7ab37b288feb43ad5c2d820c6180bb3339",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "p-value",
     "evidence": "If the resulting p-value is greater than 0.1 the power law is a plausible hypothesis for the data, otherwise it is rejected.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit p-value or inferential p notation",
     "provenance": {
      "source_path": "papers/clauset2009-power-law-data.html",
      "source_body_sha256": "fdd55be2ef634e71c4c4773765c89de3575b87479578140c6812ce0542d9e36c",
      "evidence_sha256": "c77e665539f76636f311d763eb3eb561cb4d7adf64f7a51f94a8b61695d2f491",
      "grounding_status": "not-exactly-located"
     }
    },
    {
     "term": "regression",
     "evidence": "Typically this slope is extracted by performing a least-squares linear regression on the logarithm of the histogram.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/clauset2009-power-law-data.html",
      "source_body_sha256": "fdd55be2ef634e71c4c4773765c89de3575b87479578140c6812ce0542d9e36c",
      "evidence_sha256": "cc3ed56f10f1ef0f47c4403e9615f27250d05c96c579620e666bc4d69a2f624e",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "sample-size",
     "evidence": "The method of choice for fitting parametrized models such as power-law distri- butions to observed data is the method of maximum likelihood, which provably gives accurate parameter",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/clauset2009-power-law-data.html",
      "source_body_sha256": "fdd55be2ef634e71c4c4773765c89de3575b87479578140c6812ce0542d9e36c",
      "evidence_sha256": "882d9d9c85af5f5afd194301a6b16c0bfc8eabb785f2f0a3ddbd7c857510b376",
      "grounding_status": "not-exactly-located"
     }
    },
    {
     "term": "standard-deviation",
     "evidence": "Even the largest deviations, which are exceptionally rare, are still only about a factor of two from the mean in either direction and hence the distribution can be well-characteriz",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/clauset2009-power-law-data.html",
      "source_body_sha256": "fdd55be2ef634e71c4c4773765c89de3575b87479578140c6812ce0542d9e36c",
      "evidence_sha256": "d2bcdd32bcfbd90f30ce1b6babebc5a72f495b071dff8370a1c9698e328e14af",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "standard-error",
     "evidence": "The standard error on ˆα, which is derived from the width of the likelihood maximum, is σ = ˆα 1 − √n + O(1/n) , (3.2) where the higher-order correction is positive; see Appendix B",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit standard error or SEM",
     "provenance": {
      "source_path": "papers/clauset2009-power-law-data.html",
      "source_body_sha256": "fdd55be2ef634e71c4c4773765c89de3575b87479578140c6812ce0542d9e36c",
      "evidence_sha256": "284f317818e8613cbbf6541ad1b8ed3283380313ee52378ef46175396682f0f8",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "statistical-significance",
     "evidence": "This method gives a p-value that tells us whether the observed sign of is statistically significant.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit inferential-statistical context",
     "provenance": {
      "source_path": "papers/clauset2009-power-law-data.html",
      "source_body_sha256": "fdd55be2ef634e71c4c4773765c89de3575b87479578140c6812ce0542d9e36c",
      "evidence_sha256": "e948d31358535ff5983b3ab64662a0688ab7e98258c8258216f390a1ba5a0a8c",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "variance",
     "evidence": "m this kind of fitting, provide estimates and standard errors for the slope, and calculate the fraction r2 of variance accounted for by the fitted line, which is taken as an indica",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/clauset2009-power-law-data.html",
      "source_body_sha256": "fdd55be2ef634e71c4c4773765c89de3575b87479578140c6812ce0542d9e36c",
      "evidence_sha256": "39466cbf39cc91baa240cf2f98b948984bc3ec9f4460dadcb5ca2bf384923c3f",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "arXiv-nonexclusive-distrib-1.0",
   "license_url": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
   "source_url": "https://arxiv.org/abs/0706.1062",
   "doi": "10.1137/070710111",
   "full_text_redistributable": false
  },
  {
   "id": "farhi2014-qaoa",
   "title": "A Quantum Approximate Optimization Algorithm",
   "discipline": "quantum-computing",
   "year": 2014,
   "authors": "Edward Farhi, Jeffrey Goldstone, Sam Gutmann",
   "url": "papers/farhi2014-qaoa.html",
   "key_terms": [
    "quantum",
    "algorithm",
    "optimization",
    "circuit",
    "MaxCut",
    "graphs",
    "unitary",
    "objective",
    "combinatorial"
   ],
   "phenomenon": "quantum approximate optimization (combinatorial optimization on quantum computers)",
   "claim_type": "method",
   "method": "analytic",
   "statistics": [
    {
     "term": "mean",
     "evidence": "The mean of C(z) for strings obtained in this way is Mp.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "numeric mean or average in main text",
     "provenance": {
      "source_path": "papers/farhi2014-qaoa.html",
      "source_body_sha256": "9841ebe127547dc801cb4516e5c119c55622663d04d3b6985f5ff051c366decc",
      "evidence_sha256": "b1ce34024f89fbb7279be126f96e929a3d972662d45077d4705521759a885825",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "standard-deviation",
     "evidence": "For v and p fixed we see that the standard deviation of C(z) is at most of order √m.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/farhi2014-qaoa.html",
      "source_body_sha256": "9841ebe127547dc801cb4516e5c119c55622663d04d3b6985f5ff051c366decc",
      "evidence_sha256": "d66da781602480f2cbdf5b9cdfa14fabbb6346df6b9b484d9569e7a0d40476b6",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "arXiv-nonexclusive-distrib-1.0",
   "license_url": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
   "source_url": "https://arxiv.org/abs/1411.4028",
   "doi": "",
   "full_text_redistributable": false
  },
  {
   "id": "rasp2020-weatherbench",
   "title": "WeatherBench: A Benchmark Dataset for Data-Driven Weather Forecasting",
   "discipline": "earth-science",
   "year": 2020,
   "authors": "Stephan Rasp, Peter D. Dueben, Sebastian Scher, Jonathan A. Weyn",
   "url": "papers/rasp2020-weatherbench.html",
   "key_terms": [
    "weather forecasting",
    "data-driven",
    "deep learning",
    "benchmark dataset",
    "ERA5",
    "numerical weather prediction",
    "neural networks",
    "geopotential",
    "medium-range",
    "evaluation metrics",
    "reanalysis"
   ],
   "phenomenon": "data-driven weather forecasting (machine learning benchmark)",
   "claim_type": "method",
   "method": "experiment",
   "statistics": [
    {
     "term": "correlation",
     "evidence": "The forecasts were evaluated using the root mean squared error and the anomaly correlation coefficient of Z500 and Longitudet = 0t = 5 daysa) Direct predictionb) Iterative predicti",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/rasp2020-weatherbench.html",
      "source_body_sha256": "60259c4af6282c3c64b7fd301685d6024f50464fee6a24b8717fb8de96d1a8c4",
      "evidence_sha256": "2a749136f60cd29b1e0203a4d2f01582bcebb23a96d292ae90c27a422eb130c7",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "regression",
     "evidence": "Solid lines for linear regression and CNN indicate iterative forecasts, while dots represent direct forecasts for 3 and 5 days lead time.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/rasp2020-weatherbench.html",
      "source_body_sha256": "60259c4af6282c3c64b7fd301685d6024f50464fee6a24b8717fb8de96d1a8c4",
      "evidence_sha256": "c12772e0eb51ba3396424f51ac24f69fcc6f82d66ee099927a6e2d7331ddce35",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "standard-deviation",
     "evidence": "A good first order approximation for this is the spread (ensemble standard deviation) to error (RMSE) ratio which should be one (Leutbecher and Palmer, 2008).",
     "decision": "accepted",
     "role": "substantive-mention",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/rasp2020-weatherbench.html",
      "source_body_sha256": "60259c4af6282c3c64b7fd301685d6024f50464fee6a24b8717fb8de96d1a8c4",
      "evidence_sha256": "eeea3d7f61b476f6f63564421acdea06f1eabecdd90c468b4c68da52b23ebe70",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "CC-BY-4.0",
   "license_url": "http://creativecommons.org/licenses/by/4.0/",
   "source_url": "https://arxiv.org/abs/2002.00469",
   "doi": "10.1029/2020MS002203",
   "full_text_redistributable": true
  },
  {
   "id": "collobert2011-nlp-scratch",
   "title": "Natural Language Processing (Almost) from Scratch",
   "discipline": "linguistics",
   "year": 2011,
   "authors": "Ronan Collobert, Jason Weston, Léon Bottou, Michael Karlen",
   "url": "papers/collobert2011-nlp-scratch.html",
   "key_terms": [
    "natural language processing",
    "neural network",
    "internal representations",
    "part-of-speech tagging",
    "chunking",
    "named entity recognition",
    "semantic role labeling",
    "unlabeled data",
    "benchmarks",
    "task-specific engineering"
   ],
   "phenomenon": "learned representations for NLP (avoiding hand-engineered features)",
   "claim_type": "method",
   "method": "experiment",
   "statistics": [
    {
     "term": "Bayesian",
     "evidence": "Modern speech recog- nition systems use Bayes rule to combine the outputs of an acoustic model.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/collobert2011-nlp-scratch.html",
      "source_body_sha256": "22d20e70499cb924a55120ea6c6089f5c7df2b36ec942131d5fe0cce7980f855",
      "evidence_sha256": "017ec252eca1b7970643dbe36238095dfed55a5ee39ca908c3f5c3b2cfdf7f3c",
      "grounding_status": "not-exactly-located"
     }
    },
    {
     "term": "confidence-interval",
     "evidence": "m simple n-gram models? Since such tasks certainly require high capacity models, obtaining sufficiently small confidence intervals on the test set entropy may require prohibitively",
     "decision": "accepted",
     "role": "substantive-mention",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/collobert2011-nlp-scratch.html",
      "source_body_sha256": "22d20e70499cb924a55120ea6c6089f5c7df2b36ec942131d5fe0cce7980f855",
      "evidence_sha256": "0608852a3dfd64468c8249f0be642954800e2b1e1c62b4a41e209350ca082594",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "standard-deviation",
     "evidence": "We show the standard deviation (obtained over 5 runs with different random initialization), for the architecture we picked (300 hidden units for POS, CHUNK and NER, 500 for SRL).",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/collobert2011-nlp-scratch.html",
      "source_body_sha256": "22d20e70499cb924a55120ea6c6089f5c7df2b36ec942131d5fe0cce7980f855",
      "evidence_sha256": "ff0a3afcd2f4e99d24b567177cd7b8069df8bc5dc78f774a9cf696d315aa39de",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "jmlr-open",
   "license_url": "",
   "source_url": "https://www.jmlr.org/papers/v12/collobert11a.html",
   "doi": "",
   "full_text_redistributable": false
  },
  {
   "id": "ng2017-dna2vec",
   "title": "dna2vec: Consistent Vector Representations of Variable-Length k-mers",
   "discipline": "genetics",
   "year": 2017,
   "authors": "Patrick Ng",
   "url": "papers/ng2017-dna2vec.html",
   "key_terms": [
    "k-mer",
    "vector",
    "embedding",
    "word2vec",
    "DNA",
    "similarity",
    "nucleotide",
    "representation",
    "sequence"
   ],
   "phenomenon": "word embeddings (distributed vector representations of DNA k-mers)",
   "claim_type": "method",
   "method": "experiment",
   "statistics": [
    {
     "term": "correlation",
     "evidence": "The Spearman’s rank correlation coefficient is 0.831 0.000.250.500.751.0012345678Needleman−Wunsch scoredna2vec cosine similarity Figure 2: Global alignment score distribution of ne",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/ng2017-dna2vec.html",
      "source_body_sha256": "661cde6912425cf28f962a1774b8f42226ad05251ca799f42f6ebc2c1eabb072",
      "evidence_sha256": "11bfc55fedd3eba352148329e71bf3313be1a4aaf7de26dd60a4a653a614a2a5",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "regression",
     "evidence": "Least squares support vector machine classifiers.",
     "decision": "rejected",
     "role": "non-statistical-sense",
     "reason": "lexical match does not denote the controlled statistical concept",
     "provenance": {
      "source_path": "papers/ng2017-dna2vec.html",
      "source_body_sha256": "661cde6912425cf28f962a1774b8f42226ad05251ca799f42f6ebc2c1eabb072",
      "evidence_sha256": "e22dea144a209ad0495b6ab2ea5bdf5f77605aef395b047c69cb53332fe5bc63",
      "grounding_status": "verified-reference"
     }
    }
   ],
   "license_spdx": "arXiv-nonexclusive-distrib-1.0",
   "license_url": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
   "source_url": "https://arxiv.org/abs/1701.06279",
   "doi": "",
   "full_text_redistributable": false
  },
  {
   "id": "tishby1999-information-bottleneck",
   "title": "The Information Bottleneck Method",
   "discipline": "information-theory",
   "year": 1999,
   "authors": "Naftali Tishby, Fernando C. Pereira, William Bialek",
   "url": "papers/tishby1999-information-bottleneck.html",
   "key_terms": [
    "information",
    "signal",
    "distortion",
    "compression",
    "bottleneck",
    "rate",
    "relevant",
    "quantization"
   ],
   "phenomenon": "information bottleneck (optimal tradeoff between compression and relevant information)",
   "claim_type": "method",
   "method": "analytic",
   "statistics": [
    {
     "term": "Bayesian",
     "evidence": ") exp −β \" p(y|x) log p(y|x) p(y| ̃x) # y X (16) where the distribution p(y| ̃x) in the exponent is given via Bayes’ rule and the Markov chain condition ̃X ← X ← Y , as, p(y| ̃x) =",
     "decision": "accepted",
     "role": "substantive-mention",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/tishby1999-information-bottleneck.html",
      "source_body_sha256": "8b620aab769b83793931639d86c2c6fb9d0b8869ec419005de58d9ccb77cef66",
      "evidence_sha256": "662c2f9a701aadc8fdcf3d4920a40a954c6e5df81e39416cf828c0e7a14da1e8",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "mean",
     "evidence": "Equation (28) also have the interpretation of a weighted average of the data conditional distributions that contribute to the representative ̃x.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/tishby1999-information-bottleneck.html",
      "source_body_sha256": "8b620aab769b83793931639d86c2c6fb9d0b8869ec419005de58d9ccb77cef66",
      "evidence_sha256": "cde0144a89d0a6a6078ca9a071c837d7c593032532a09d0691af7a79f335256e",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "arxiv-open",
   "license_url": "",
   "source_url": "https://arxiv.org/abs/physics/0004057",
   "doi": "",
   "full_text_redistributable": false
  },
  {
   "id": "bouchaud2008-market-impact",
   "title": "How Markets Slowly Digest Changes in Supply and Demand",
   "discipline": "quantitative-finance",
   "year": 2009,
   "authors": "Jean-Philippe Bouchaud, J. Doyne Farmer, Fabrizio Lillo",
   "url": "papers/bouchaud2008-market-impact.html",
   "key_terms": [
    "market impact",
    "supply and demand",
    "order flow",
    "liquidity",
    "market efficiency",
    "price formation",
    "long-memory",
    "bid-ask spread",
    "volatility",
    "microstructure"
   ],
   "phenomenon": "market impact of trades (concave impact + long-memory / power-law autocorrelation of order flow)",
   "claim_type": "mechanism",
   "method": "review",
   "statistics": [
    {
     "term": "Bayesian",
     "evidence": "But as trading occurs, either at the bid or at the ask, the market maker updates in a Bayesian way the value of δ+ = 1 − δ−: trades at the ask increase the value of δ+, while trade",
     "decision": "accepted",
     "role": "substantive-mention",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/bouchaud2008-market-impact.html",
      "source_body_sha256": "57e1888db24269d79c9f2f672385c9fd2d6b41baee644516bf756fd0447dbc11",
      "evidence_sha256": "2f82a0cee4f648512a4b0f7eccf5368e937945b058f019a771d1880b7cb04c1f",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "confidence-interval",
     "evidence": "Unfortunately there are still not good statistical methods for assigning confidence intervals for exponents of power laws, particularly when the observations have long-memory, but ",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/bouchaud2008-market-impact.html",
      "source_body_sha256": "57e1888db24269d79c9f2f672385c9fd2d6b41baee644516bf756fd0447dbc11",
      "evidence_sha256": "eef387437c230dd75351275e8437acafc3a5b03744945077961e719322131a63",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "correlation",
     "evidence": "− j) + (cid:88) j>0 [G0((cid:96) + j) − G0(j)]2 + 2∆((cid:96)) + Σ2(cid:96), (6.16) where ∆((cid:96)) is the correlation induced contribution: ∆((cid:96)) = (cid:88) G0((cid:96) − ",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/bouchaud2008-market-impact.html",
      "source_body_sha256": "57e1888db24269d79c9f2f672385c9fd2d6b41baee644516bf756fd0447dbc11",
      "evidence_sha256": "d133096921554b5df56baf61ae9b6269feda8bc1cdcfe3a173ba52f64e672d1e",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "hypothesis-testing",
     "evidence": "have used membership codes of the London Stock Exchange to test the hypothesis of the theory presented in Lillo et al.",
     "decision": "rejected",
     "role": "non-statistical-sense",
     "reason": "lexical match does not denote the controlled statistical concept",
     "provenance": {
      "source_path": "papers/bouchaud2008-market-impact.html",
      "source_body_sha256": "57e1888db24269d79c9f2f672385c9fd2d6b41baee644516bf756fd0447dbc11",
      "evidence_sha256": "6cb7ce615196edbd54f3f9b5d8e735c815a7ba270badf54e96ccaab020fa0ea0",
      "grounding_status": "not-exactly-located"
     }
    },
    {
     "term": "mean",
     "evidence": "We have plotted the average value of the return conditioned to be in the direction of the predictor, r+, and the average return when the sign of the predictor is wrong, r−.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/bouchaud2008-market-impact.html",
      "source_body_sha256": "57e1888db24269d79c9f2f672385c9fd2d6b41baee644516bf756fd0447dbc11",
      "evidence_sha256": "4136cf6cf7f97d4bc7584e0a687665fc9b135f9ad7cb7ca2aa47fab7f603268a",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "regression",
     "evidence": "At a daily scale he finds he gets the best linear regression against contemporary daily returns with ψ ≈ 0.5.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/bouchaud2008-market-impact.html",
      "source_body_sha256": "57e1888db24269d79c9f2f672385c9fd2d6b41baee644516bf756fd0447dbc11",
      "evidence_sha256": "9183995b2ba3c0d88de42b9acdbfe444f5eb9a080b65fe60b70f0cc77abe1f73",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "standard-deviation",
     "evidence": "Analysis of price impact makes INFORMATION, LIQUIDITY & EFFICIENCY 13 it clear that the standard deviation of impacts is very large compared to their mean, suggesting that this is ",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/bouchaud2008-market-impact.html",
      "source_body_sha256": "57e1888db24269d79c9f2f672385c9fd2d6b41baee644516bf756fd0447dbc11",
      "evidence_sha256": "badb1eb86cff971a770c6940c621c0c7ce018302460864645078056d69704f53",
      "grounding_status": "not-exactly-located"
     }
    },
    {
     "term": "standard-error",
     "evidence": "Error bars are standard errors.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit standard error or SEM",
     "provenance": {
      "source_path": "papers/bouchaud2008-market-impact.html",
      "source_body_sha256": "57e1888db24269d79c9f2f672385c9fd2d6b41baee644516bf756fd0447dbc11",
      "evidence_sha256": "6302b9df973f71ca1b463c60699f868f909ac0cc0d6ee720c41e718d8b735a76",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "statistical-significance",
     "evidence": "It is observed in all the stock markets studied so far at very high levels of statistical significance.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit inferential-statistical context",
     "provenance": {
      "source_path": "papers/bouchaud2008-market-impact.html",
      "source_body_sha256": "57e1888db24269d79c9f2f672385c9fd2d6b41baee644516bf756fd0447dbc11",
      "evidence_sha256": "1574cf20d42ad21a72d64ea555a27a8e2ceeaa5a0af44b17bebeaee4c57cd8af",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "variance",
     "evidence": "(6.13) The lagged impact and the return variance would then be given by: R(cid:96) = α(cid:96)−1E[f ]; V(cid:96) = 2E[f 2] 1 − α(cid:96) 1 − α2 + Σ2(cid:96), (6.14) i.e.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/bouchaud2008-market-impact.html",
      "source_body_sha256": "57e1888db24269d79c9f2f672385c9fd2d6b41baee644516bf756fd0447dbc11",
      "evidence_sha256": "b68860874ae0c3dc154b634960bfa7d5efcb13d4d173b892d667e1433768de43",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "arXiv-nonexclusive-distrib-1.0",
   "license_url": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
   "source_url": "https://arxiv.org/abs/0809.0822",
   "doi": "",
   "full_text_redistributable": false
  },
  {
   "id": "abbott2017-gw170817",
   "title": "GW170817: Observation of Gravitational Waves from a Binary Neutron Star Inspiral",
   "discipline": "astronomy",
   "year": 2017,
   "authors": "LIGO Scientific Collaboration, Virgo Collaboration",
   "url": "papers/abbott2017-gw170817.html",
   "key_terms": [
    "gravitational waves",
    "binary neutron star",
    "inspiral",
    "LIGO",
    "Virgo",
    "signal-to-noise ratio",
    "chirp mass",
    "gamma-ray burst",
    "neutron star merger",
    "luminosity distance",
    "coalescence"
   ],
   "phenomenon": "gravitational waves (binary neutron star merger; multi-messenger observation)",
   "claim_type": "existence",
   "method": "observation",
   "statistics": [
    {
     "term": "Bayesian",
     "evidence": "We perform a Bayesian analysis in the frequency range 30–2048 Hz that includes the effects of the 1σ calibration uncertainties on t",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/abbott2017-gw170817.html",
      "source_body_sha256": "c58fb42f86cfea36d270d2460a9f1931328607a266d3243c85c2533c09ff514c",
      "evidence_sha256": "1fa71519f30fae3dbabdfe403f58613b233a2bf66ec454b7d823b33449efe28f",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "correlation",
     "evidence": "The individual masses are in the broad range of 0.86 to 2.26 M⊙, due to correlations between their uncertainties.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/abbott2017-gw170817.html",
      "source_body_sha256": "c58fb42f86cfea36d270d2460a9f1931328607a266d3243c85c2533c09ff514c",
      "evidence_sha256": "d82f3d78dac2f6168f11d0f2c69502c5d6a154a79960af7aa6c31235ce64100a",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "null-hypothesis",
     "evidence": "The estimated Hubble flow velocity near NGC 4993 of 3017 (cid:2) 166 km s−1 [141] provides a redshift, which in a flat cosmology with H0 1⁄4 67.90 (cid:2) 0.55 km s−1 Mpc−1 [90], c",
     "decision": "rejected",
     "role": "symbol-collision",
     "reason": "H0 denotes a Hubble parameter, Hamiltonian, operator, or generating function rather than a null hypothesis",
     "provenance": {
      "source_path": "papers/abbott2017-gw170817.html",
      "source_body_sha256": "c58fb42f86cfea36d270d2460a9f1931328607a266d3243c85c2533c09ff514c",
      "evidence_sha256": "e31e460e4de41ef667f9cb4389f5aba6c8877bcf74aef9c6cec8794f659a600a",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "CC-BY-4.0",
   "license_url": "http://creativecommons.org/licenses/by/4.0/",
   "source_url": "https://arxiv.org/abs/1710.05832",
   "doi": "10.1103/PhysRevLett.119.161101",
   "full_text_redistributable": true
  },
  {
   "id": "kitaev2003-anyons",
   "title": "Fault-Tolerant Quantum Computation by Anyons",
   "discipline": "physics",
   "year": 2003,
   "authors": "Alexei Yu. Kitaev",
   "url": "papers/kitaev2003-anyons.html",
   "key_terms": [
    "quantum",
    "excitations",
    "topological",
    "fault-tolerant",
    "unitary",
    "stabilizer",
    "degenerate",
    "braiding",
    "anyons",
    "error"
   ],
   "phenomenon": "fault-tolerant (topological) quantum computation by anyons",
   "claim_type": "method",
   "method": "analytic",
   "statistics": [
    {
     "term": "mean",
     "evidence": "If the gauge symmetry is broken then there is a nonvanishing vacuum average of the Higgs field, φ(s) = = 0.",
     "decision": "rejected",
     "role": "non-statistical-sense",
     "reason": "lexical match does not denote the controlled statistical concept",
     "provenance": {
      "source_path": "papers/kitaev2003-anyons.html",
      "source_body_sha256": "53a0fe0c0cd8fdf25a54aefaa3c7d0d7846e22aea4e3ba61cfb24ed0cab4ac45",
      "evidence_sha256": "a9b82b81309e36bd10751be2a7fbab9344a3b86065e03d2b34479ef40256c3e9",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "null-hypothesis",
     "evidence": "In particular, 0 = U H0 U † = H ′ 0 ≡ − X s σx s − X p σz p (8) 1, Qp ≡ H ′ In the field theory language, the vertex variables vs (or the operators σz s ) are a Higgs field.",
     "decision": "rejected",
     "role": "symbol-collision",
     "reason": "H0 denotes a Hubble parameter, Hamiltonian, operator, or generating function rather than a null hypothesis",
     "provenance": {
      "source_path": "papers/kitaev2003-anyons.html",
      "source_body_sha256": "53a0fe0c0cd8fdf25a54aefaa3c7d0d7846e22aea4e3ba61cfb24ed0cab4ac45",
      "evidence_sha256": "83f315c0e774808d0d80c697116d3d46b453a9fc136db8696ec19b1e4f96ae5f",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "standard-error",
     "evidence": "This process is quite similar to the standard error correction procedure for the repetition code.",
     "decision": "rejected",
     "role": "non-statistical-sense",
     "reason": "lexical match does not denote the controlled statistical concept",
     "provenance": {
      "source_path": "papers/kitaev2003-anyons.html",
      "source_body_sha256": "53a0fe0c0cd8fdf25a54aefaa3c7d0d7846e22aea4e3ba61cfb24ed0cab4ac45",
      "evidence_sha256": "0f95840fa397cca917c3ebeec3136836e8943b59a2011614415826d52e375b53",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "arxiv-open",
   "license_url": "",
   "source_url": "https://arxiv.org/abs/quant-ph/9707021",
   "doi": "10.1016/S0003-4916(02)00018-0",
   "full_text_redistributable": false
  },
  {
   "id": "litjens2017-medical-imaging-survey",
   "title": "A Survey on Deep Learning in Medical Image Analysis",
   "discipline": "medicine",
   "year": 2017,
   "authors": "Geert Litjens, Thijs Kooi, Babak Ehteshami Bejnordi, Arnaud Arindra Adiyoso Setio",
   "url": "papers/litjens2017-medical-imaging-survey.html",
   "key_terms": [
    "deep learning",
    "convolutional networks",
    "medical image analysis",
    "image classification",
    "segmentation",
    "object detection",
    "feature extraction",
    "pattern recognition",
    "supervised",
    "medical imaging"
   ],
   "phenomenon": "deep learning in medical image analysis (survey of the field)",
   "claim_type": "existence",
   "method": "review",
   "statistics": [
    {
     "term": "Bayesian",
     "evidence": "Researchers have also looked at Bayesian methods for hyper-parameter opti- mization (Snoek et al., 2012), but this has not been ap- plied in medical image analysis as far as we are",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/litjens2017-medical-imaging-survey.html",
      "source_body_sha256": "1d4c19ecd99067d5f827ee3d6c359e61eeaed4c3c166c6450e7388ccbc9e745c",
      "evidence_sha256": "324565d36d48e0797e3e51ccefec09db7556f27cef63b3503d0a78c6d3b75fcc",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "regression",
     "evidence": "traction, HMM as a generative model on top AD/MCI/HC classification; CNN on sparse representations created by regression models AD/MCI/HC classification; Multi-modal stacked deep p",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/litjens2017-medical-imaging-survey.html",
      "source_body_sha256": "1d4c19ecd99067d5f827ee3d6c359e61eeaed4c3c166c6450e7388ccbc9e745c",
      "evidence_sha256": "eb29749007c2365dbb046a2e5af517069256163f1f1760bb763d5f9c4eff6c10",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "statistical-significance",
     "evidence": "They showed that their approach has significantly higher reg- istration success rates than using traditional - purely in- tensity based - registration methods.",
     "decision": "rejected",
     "role": "lexical-significance",
     "reason": "significant/significantly is used without inferential-statistical meaning",
     "provenance": {
      "source_path": "papers/litjens2017-medical-imaging-survey.html",
      "source_body_sha256": "1d4c19ecd99067d5f827ee3d6c359e61eeaed4c3c166c6450e7388ccbc9e745c",
      "evidence_sha256": "fba067b70ce7f34709ea1038e203e90731f43ec544b3680a3e3beb1c326af914",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "arXiv-nonexclusive-distrib-1.0",
   "license_url": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
   "source_url": "https://arxiv.org/abs/1702.05747",
   "doi": "10.1016/j.media.2017.07.005",
   "full_text_redistributable": false
  },
  {
   "id": "lake2018-generalization-systematicity",
   "title": "Generalization without Systematicity: On the Compositional Skills of Sequence-to-Sequence Recurrent Networks",
   "discipline": "cognitive-science",
   "year": 2018,
   "authors": "Brenden M. Lake, Marco Baroni",
   "url": "papers/lake2018-generalization-systematicity.html",
   "key_terms": [
    "compositional",
    "generalization",
    "systematicity",
    "RNNs",
    "SCAN",
    "commands",
    "recurrent",
    "zero-shot",
    "networks"
   ],
   "phenomenon": "systematic compositional generalization (in neural sequence models)",
   "claim_type": "critique",
   "method": "experiment",
   "statistics": [
    {
     "term": "Bayesian",
     "evidence": "Efficient Gradient-Based Inference through Transformations between Bayes Nets and Neural Nets.",
     "decision": "rejected",
     "role": "bibliography",
     "reason": "occurrence is confined to a citation/reference rather than the article body",
     "provenance": {
      "source_path": "papers/lake2018-generalization-systematicity.html",
      "source_body_sha256": "b4f73f8241a89d38ebd7ca5a13b5ad57a1e37801c5f155e5ffbe0f9a71e8612f",
      "evidence_sha256": "f9ef6b3fe8d6b4d67bc7b065214769cfc35af48ce3619c7d4cadea81a919f798",
      "grounding_status": "verified-reference"
     }
    },
    {
     "term": "mean",
     "evidence": "n-less test command that does not also occur in the training split, and the frequency of occurrence of such commands in the training set is at a non-negligible average value of 4 t",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/lake2018-generalization-systematicity.html",
      "source_body_sha256": "b4f73f8241a89d38ebd7ca5a13b5ad57a1e37801c5f155e5ffbe0f9a71e8612f",
      "evidence_sha256": "31a4f2b05752c757667abf52c21121ddb17aff5531cef39821d14cfb1a28348b",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "standard-error",
     "evidence": "Each bar shows the mean over 5 training runs with corresponding ±1 SEM.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit standard error or SEM",
     "provenance": {
      "source_path": "papers/lake2018-generalization-systematicity.html",
      "source_body_sha256": "b4f73f8241a89d38ebd7ca5a13b5ad57a1e37801c5f155e5ffbe0f9a71e8612f",
      "evidence_sha256": "78b8b249b45fae1751c042be0178a0dd93fe2b615ed6e89e217a7d6d83ad4f82",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "arXiv-nonexclusive-distrib-1.0",
   "license_url": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
   "source_url": "https://arxiv.org/abs/1711.00350",
   "doi": "",
   "full_text_redistributable": false
  },
  {
   "id": "hardt2016-equality-opportunity",
   "title": "Equality of Opportunity in Supervised Learning",
   "discipline": "law",
   "year": 2016,
   "authors": "Moritz Hardt, Eric Price, Nathan Srebro",
   "url": "papers/hardt2016-equality-opportunity.html",
   "key_terms": [
    "discrimination",
    "supervised learning",
    "protected attribute",
    "fairness",
    "demographic parity",
    "predictor",
    "equality of opportunity",
    "oblivious",
    "sensitive attribute",
    "classification"
   ],
   "phenomenon": "algorithmic fairness (equalized odds / equal opportunity criterion)",
   "claim_type": "method",
   "method": "analytic",
   "statistics": [
    {
     "term": "Bayesian",
     "evidence": "Moreover, we quantify the loss that follows from imposing our non-discrimination condition in case the score we start from deviates from Bayesian optimality.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/hardt2016-equality-opportunity.html",
      "source_body_sha256": "ee3bea023560924353a3d2e94be7e5dc00f88465d9c7082fa21126fc16781d75",
      "evidence_sha256": "67827c3c8353fbd10f55ad75193fe625b57ddd75d0da43ebd7b7f5b7d33c6204",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "correlation",
     "evidence": "If A and X3, and thus A and Y , are positively correlated, then (cid:101)R would depend inversely on A (see numerical construction below), introducing a form of “corre",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/hardt2016-equality-opportunity.html",
      "source_body_sha256": "ee3bea023560924353a3d2e94be7e5dc00f88465d9c7082fa21126fc16781d75",
      "evidence_sha256": "b01ce2222d6824603329274a589712015dfe9ad99da2e268ecf9211c0c535b3b",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "arXiv-nonexclusive-distrib-1.0",
   "license_url": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
   "source_url": "https://arxiv.org/abs/1610.02413",
   "doi": "",
   "full_text_redistributable": false
  },
  {
   "id": "ribeiro2016-lime",
   "title": "Why Should I Trust You? Explaining the Predictions of Any Classifier",
   "discipline": "statistics",
   "year": 2016,
   "authors": "Marco Tulio Ribeiro, Sameer Singh, Carlos Guestrin",
   "url": "papers/ribeiro2016-lime.html",
   "key_terms": [
    "explanation",
    "prediction",
    "model",
    "interpretable",
    "classifier",
    "trust",
    "faithful",
    "features",
    "LIME"
   ],
   "phenomenon": "interpretability / explainability of machine-learning models (explaining black-box predictions)",
   "claim_type": "method",
   "method": "experiment",
   "statistics": [
    {
     "term": "Bayesian",
     "evidence": "Machine learning practitioners may be able to interpret small Bayesian networks, but laymen may be more comfortable with a small number of weighted features as an explanation.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/ribeiro2016-lime.html",
      "source_body_sha256": "6d7e6be6d1ea00df537696ff5f7876c5ac99e5c94a6de055c25af17623ac6670",
      "evidence_sha256": "59dfeffd125971ee9af9f511e574569365098cb2484f6a96856cab3fed112261",
      "grounding_status": "not-exactly-located"
     }
    },
    {
     "term": "mean",
     "evidence": "Each subject took an average of 3.6 minutes per round of cleaning, resulting in just under 11 minutes to produce a classifier that generalizes much better to real world data.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/ribeiro2016-lime.html",
      "source_body_sha256": "6d7e6be6d1ea00df537696ff5f7876c5ac99e5c94a6de055c25af17623ac6670",
      "evidence_sha256": "35018fe7f63636150e9c71d2ce7061283b348238c1f6a9e90082a3fb81f77a2c",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "p-value",
     "evidence": "The top 3 classes predicted are “Electric Guitar” (p = 0.32), “Acoustic guitar” (p = 0.24) and “Labrador” (p = 0.21) 4.",
     "decision": "rejected",
     "role": "non-statistical-sense",
     "reason": "lexical match does not denote the controlled statistical concept",
     "provenance": {
      "source_path": "papers/ribeiro2016-lime.html",
      "source_body_sha256": "6d7e6be6d1ea00df537696ff5f7876c5ac99e5c94a6de055c25af17623ac6670",
      "evidence_sha256": "aa99c26e86a3fbf1cdb0085031033353c6a5ed35bd87c69753706b0d57727e8a",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "regression",
     "evidence": "We train decision trees (DT), logistic regression with L2 regularization (LR), nearest neighbors (NN), and support vector machines with RBF kernel (SVM), all using bag of words as ",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/ribeiro2016-lime.html",
      "source_body_sha256": "6d7e6be6d1ea00df537696ff5f7876c5ac99e5c94a6de055c25af17623ac6670",
      "evidence_sha256": "135d15adf615aeed27da03a105a6f25da69a015c6a8e6f72434f011c71c0cf4c",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "sample-size",
     "evidence": "Although our sample size is small, this experiment demonstrates the utility of explaining indi- vidual predictions for getting insights into classifiers knowing when not to trust t",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/ribeiro2016-lime.html",
      "source_body_sha256": "6d7e6be6d1ea00df537696ff5f7876c5ac99e5c94a6de055c25af17623ac6670",
      "evidence_sha256": "af34f4f51dee98261de05f8b3b5b1b1e33655a67c65eb94002a4d5274127bb3b",
      "grounding_status": "not-exactly-located"
     }
    },
    {
     "term": "standard-error",
     "evidence": "Averages and standard errors from 800 runs.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit standard error or SEM",
     "provenance": {
      "source_path": "papers/ribeiro2016-lime.html",
      "source_body_sha256": "6d7e6be6d1ea00df537696ff5f7876c5ac99e5c94a6de055c25af17623ac6670",
      "evidence_sha256": "ac1437dc7b2d7f51dfe33d77aa82de09f363cf66d10d94211805809cf6ee8a63",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "statistical-significance",
     "evidence": "However, real-world data is often significantly different, and further, the evaluation metric may not be indicative of the product’s goal.",
     "decision": "rejected",
     "role": "lexical-significance",
     "reason": "significant/significantly is used without inferential-statistical meaning",
     "provenance": {
      "source_path": "papers/ribeiro2016-lime.html",
      "source_body_sha256": "6d7e6be6d1ea00df537696ff5f7876c5ac99e5c94a6de055c25af17623ac6670",
      "evidence_sha256": "4825f9c01020e9f1c552231f785361e2b0fd21ccf5f9b8ba240df210a39c071b",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "arXiv-nonexclusive-distrib-1.0",
   "license_url": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
   "source_url": "https://arxiv.org/abs/1602.04938",
   "doi": "",
   "full_text_redistributable": false
  },
  {
   "id": "hoffman2013-svi",
   "title": "Stochastic Variational Inference",
   "discipline": "statistics",
   "year": 2013,
   "authors": "Matthew D. Hoffman, David M. Blei, Chong Wang, John Paisley",
   "url": "papers/hoffman2013-svi.html",
   "key_terms": [
    "stochastic variational inference",
    "posterior",
    "probabilistic models",
    "topic models",
    "latent Dirichlet allocation",
    "Bayesian",
    "massive data",
    "scalable",
    "probabilistic modeling",
    "variational inference"
   ],
   "phenomenon": "scalable Bayesian inference (stochastic variational inference for massive data)",
   "claim_type": "method",
   "method": "analytic",
   "statistics": [
    {
     "term": "Bayesian",
     "evidence": "(We also show that the Bayesian nonparametric topic model outperforms its parametric counterpart.) Stochastic variational inference lets us apply complex Bayesian models to massive",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/hoffman2013-svi.html",
      "source_body_sha256": "8bfa51984d4aa26816602792e9d85c26d2d7095a355251f68beff1e6bdf6ea37",
      "evidence_sha256": "2a453feb663ccc333f63bf65594e4283f8ce8f023c01dfeb74ccf49dd3507236",
      "grounding_status": "not-exactly-located"
     }
    },
    {
     "term": "correlation",
     "evidence": "Nonconjugate models can capture correlations between topics.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/hoffman2013-svi.html",
      "source_body_sha256": "8bfa51984d4aa26816602792e9d85c26d2d7095a355251f68beff1e6bdf6ea37",
      "evidence_sha256": "6c9232f3205a49097ba74f72c96d705a29d0ec9f71cb73226e84bc2a35e921dc",
      "grounding_status": "not-exactly-located"
     }
    },
    {
     "term": "mean",
     "evidence": "parameters using classical coordinate ascent updates where the sampled data point is repeated N times; finally, we set the new global parameters to a weighted average of the old es",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "manual audit: unambiguous statistical/probabilistic sense",
     "provenance": {
      "source_path": "papers/hoffman2013-svi.html",
      "source_body_sha256": "8bfa51984d4aa26816602792e9d85c26d2d7095a355251f68beff1e6bdf6ea37",
      "evidence_sha256": "1aa6038aacc359b84cab8d29820d0aa8c3012da3175044d4faba60cf06a1a62d",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "regression",
     "evidence": "any variants) (Kalman, 1960; Fox et al., 2011a), factorial models (Ghahramani and Jordan, 1997), hierarchical linear regression models (Gelman and Hill, 2007), hierarchical probit ",
     "decision": "accepted",
     "role": "substantive-mention",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/hoffman2013-svi.html",
      "source_body_sha256": "8bfa51984d4aa26816602792e9d85c26d2d7095a355251f68beff1e6bdf6ea37",
      "evidence_sha256": "3a933939a73e78246001a31b831e715d4d80e3cd576fedd22b52be7640f90ba0",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "sample-size",
     "evidence": "To update the global topic parameters, we again form intermediate topics with the sampled document’s optimized local parameters, ˆλk = η + D T i=1 Eq[ck di] N n=1 Eq[zi dn]wdn.",
     "decision": "rejected",
     "role": "non-statistical-sense",
     "reason": "lexical match does not denote the controlled statistical concept",
     "provenance": {
      "source_path": "papers/hoffman2013-svi.html",
      "source_body_sha256": "8bfa51984d4aa26816602792e9d85c26d2d7095a355251f68beff1e6bdf6ea37",
      "evidence_sha256": "cf6bc199538f16d5b891e5f61f87d851f8a22dbea017c972366139e7421930f7",
      "grounding_status": "verified-main-text"
     }
    },
    {
     "term": "variance",
     "evidence": "The global variables are the mixture proportions and the means and variances of the mixture components; the local variable zn is the hidden cluster label for the nth observation xn",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/hoffman2013-svi.html",
      "source_body_sha256": "8bfa51984d4aa26816602792e9d85c26d2d7095a355251f68beff1e6bdf6ea37",
      "evidence_sha256": "94d54fa078e5f90f1ab17535b0b0d53119da4b1a22a6d68fc28e52f4d8cc1404",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "arXiv-nonexclusive-distrib-1.0",
   "license_url": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
   "source_url": "https://arxiv.org/abs/1206.7051",
   "doi": "",
   "full_text_redistributable": false
  },
  {
   "id": "nguyen2020-text-as-data",
   "title": "How We Do Things with Words: Analyzing Text as Social and Cultural Data",
   "discipline": "digital-humanities",
   "year": 2020,
   "authors": "Dong Nguyen, Maria Liakata, Simon DeDeo, Jacob Eisenstein",
   "url": "papers/nguyen2020-text-as-data.html",
   "key_terms": [
    "text",
    "computational",
    "social",
    "humanities",
    "annotation",
    "validity",
    "measurement",
    "interpretation",
    "corpus"
   ],
   "phenomenon": "computational text analysis of social and cultural data (text-as-data; humanities-computation bridge)",
   "claim_type": "critique",
   "method": "review",
   "statistics": [
    {
     "term": "hypothesis-testing",
     "evidence": "For example, Nguyen and Eisenstein [2017] perform exploration in the context of hypothesis testing by using latitude and longitude coordinates — the original metadata attached to g",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "explicit hypothesis-testing language",
     "provenance": {
      "source_path": "papers/nguyen2020-text-as-data.html",
      "source_body_sha256": "f74c0204e016ba27a80d013285bb3a4c15d59d9b3f30061e27f82cdad047e882",
      "evidence_sha256": "747abe76252380dd70c5650e7bdbcddbfea724a5938a1db9dbf8e536ee52c841",
      "grounding_status": "not-exactly-located"
     }
    },
    {
     "term": "sample-size",
     "evidence": "Expanding the collection beyond the documents that we would immediately think of has the beneficial effect of increasing our sample size.",
     "decision": "accepted",
     "role": "focal-use",
     "reason": "unambiguous statistical/probabilistic term in main text",
     "provenance": {
      "source_path": "papers/nguyen2020-text-as-data.html",
      "source_body_sha256": "f74c0204e016ba27a80d013285bb3a4c15d59d9b3f30061e27f82cdad047e882",
      "evidence_sha256": "e7954db0b8a4ef66ea0062afba701cc504380bbc726b817df68e2f71c9daa77f",
      "grounding_status": "verified-main-text"
     }
    }
   ],
   "license_spdx": "CC-BY-4.0",
   "license_url": "https://creativecommons.org/licenses/by/4.0/",
   "source_url": "https://arxiv.org/abs/1907.01468",
   "doi": "10.3389/frai.2020.00062",
   "full_text_redistributable": true
  }
 ],
 "rights_manifest": "DATA/rights_manifest.json"
}