phenomenon · claim-type · cross-domain-transfer · mechanism · claim-relation · replication-status · limitation — grounded in the full text| phenomenon | medical data sitting in silos that machine learning cannot reach— “Another reason why data sharing is not systematic in healthcare is that collecting, curating, and maintaining a high-quality data set takes considerable time, effort, and expense. Consequently such data sets may have significant business value, making it less likely that they will be freely shared.” | |
| claim-type | prediction— “Not all technical questions have been answered yet and FL will certainly be an active research area throughout the next decade” | |
| method — evidence | — “Recent research has shown that models trained by FL can achieve performance levels comparable to ones trained on centrally hosted data sets and superior to models that only see isolated single-institutional data16,17.” | |
| object of study — evidence | — “Federated learning (FL)9–11 is a learning paradigm seeking to address the problem of data governance and privacy by training algorithms collaboratively without exchanging the data itself.” | |
| research question — evidence | — “By enabling multiple parties to train collaboratively without the need to exchange or centralise data sets, FL neatly addresses issues related to egress of sensitive medical data.” | |
| contribution — evidence | — “However, FL still requires rigorous technical consideration to ensure that the algorithm is proceeding optimally without compromising safety or patient privacy.” | |
| mechanism | MOVE THE MODEL, NOT THE DATA: training happens locally inside each institution and only model characteristics leave the firewall, so the consensus model is built without the patient records ever being pooled— “FL enables gaining insights collaboratively, e.g., in the form of a consensus model, without moving patient data beyond the firewalls of the institutions in which they reside. Instead, the ML process occurs locally at each participating institution and only model characteristics (e.g., parameters, gradients) are transferred as depicted in… | mechanisms[0] |
| mechanism | WHY ANONYMISATION CANNOT BE THE ANSWER: identity survives the removal of the obvious fields - a face is reconstructable from a CT scan, and a handful of record elements permits reidentification— “Even if data anonymisation could bypass these limitations, it is now well understood that removing metadata such as patient name or date of birth is often not enough to preserve privacy7.” | mechanisms[1] |
| mechanism | DATA IS WITHHELD FOR COMMERCIAL REASONS AS WELL AS LEGAL ONES: curating a high-quality set costs time and money, which gives it business value and makes sharing less likely - a motive no privacy technology addresses— “Another reason why data sharing is not systematic in healthcare is that collecting, curating, and maintaining a high-quality data set takes considerable time, effort, and expense. Consequently such data sets may have significant business value, making it less likely that they will be freely shared.” | mechanisms[2] |
| sense | anonymisation - TREATED HERE AS INSUFFICIENT BY CONSTRUCTION, not as imperfect in practice: removing name and date of birth does not preserve privacy, a face can be reconstructed from a CT scan, and genomic data is as unique as a fingerprint— “Even if data anonymisation could bypass these limitations, it is now well understood that removing metadata such as patient name or date of birth is often not enough to preserve privacy7.” | senses[0] |
| sense | Federated learning - collaborative model training without exchanging or centralising the parties' data sets— “Federated learning definition FL is a learning paradigm in which multiple parties train collaboratively without the need to exchange or centralise data sets. A general formulation of FL reads as follows: Let L denote a global loss function obtained via a weighted combination of K local losses fLkgK k¼1, computed from private data Xk, which is… | senses[1] |
| sense | Aggregation-server FL - nodes receive a global model, train locally, and return partial models for central aggregation— “Fig. 1 Example federated learning (FL) workflows and difference to learning on a Centralised Data Lake. a FL aggregation server—the typical FL workflow in which a federation of training nodes receive the global model, resubmit their partially trained models to a central server intermittently for aggregation and then continue training on the… | senses[2] |
| sense | Peer-to-peer FL - each node exchanges partial models with peers and performs its own aggregation— “Fig. 1 Example federated learning (FL) workflows and difference to learning on a Centralised Data Lake. a FL aggregation server—the typical FL workflow in which a federation of training nodes receive the global model, resubmit their partially trained models… | senses[3] |
| sense | Centralised training - sites donate data to a shared lake for extraction and independent local training | senses[4] |
| sense | FL topology - centralised, decentralised, or hierarchical communication architecture Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | senses[5] |
| sense | FL compute plan - trajectory followed by a model across participating partners Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | senses[6] |
| sense | Data heterogeneity - non-IID cross-institutional distributions that challenge federated aggregation and local optimality Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | senses[7] |
| sense | Privacy-performance trade-off - stronger federated protections may affect final-model accuracy Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | senses[8] |
| sense | Non-trusted federation - collaboration needing encryption, authentication, traceability, verification, integrity, and adversarial defenses Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | senses[9] |
| sense | Federated information leakage - model changes exposed to multiple parties create reverse-engineering risk even without raw-data sharing Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | senses[10] |
| sense | Honest-broker administration - trusted intermediary controls the federation and abstracts internal mechanisms from clients Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | senses[11] |
| sense | Peer-to-peer administration - sites interact directly with no gatekeeper and must synchronise protocol changes Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | senses[12] |
| audit | Federated learning —distinct-from→ anonymisation — Audit 2026-09-28: DOUBTFUL - read against its span and kept, but the span describes federated learning's privacy aim; distinct-from 'anonymisation' is loose. Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concept_edges[0] |
| weight | Federated learning —distinct-from→ anonymisation — 0.8 Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concept_edges[0] |
| effect direction | without access to sufficient data —causes→ transition from research to clinical practice — eliminates Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concept_edges[1] |
| note | without access to sufficient data —causes→ transition from research to clinical practice — The paper says insufficient access prevents ML from making the research-to-clinic transition. Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concept_edges[1] |
| note | Centralised training —distinct-from→ FL aggregation server — The figure explicitly presents the difference between federated workflows and the non-FL centralised-data-lake workflow. Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concept_edges[2] |
| audit | learning from encrypted data —causes→ privacy — Audit 2026-09-28: DOUBTFUL - read against its full span and kept, but the span says learning from encrypted data HAS BEEN PROPOSED to further enhance privacy; hedged. Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concept_edges[3] |
| effect direction | learning from encrypted data —causes→ privacy — increases Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concept_edges[3] |
| evidence status note | learning from encrypted data —causes→ privacy — THE QUOTE CONTINUES A SENTENCE THAT THIS COPY CUTS IN TWO - a figure caption (the FL compute plans panels) stands between it and the words it follows, which are not adjacent to it in the extraction. Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concept_edges[3] |
| note | learning from encrypted data —causes→ privacy — The same proposed-mechanism hedge applies, so paper weight is 0.7 while confidence remains 0.95. Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concept_edges[3] |
| weight | learning from encrypted data —causes→ privacy — 0.7 Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concept_edges[3] |
| note | FL —is-a→ learning paradigm — The paper directly defines FL as a learning paradigm. Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concept_edges[4] |
| audit | global optimal solution —distinct-from→ optimal for an individual local participant — Audit 2026-09-28: DOUBTFUL - read against its full span and kept, but the span says the global optimum MAY not be optimal for a local participant; hedged. Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concept_edges[5] |
| note | global optimal solution —distinct-from→ optimal for an individual local participant — The paper says the global optimum may not be locally optimal; weight 0.7 preserves both modal hedges. Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concept_edges[5] |
| weight | global optimal solution —distinct-from→ optimal for an individual local participant — 0.7 Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concept_edges[5] |
| effect direction | training process is exposed to multiple parties —causes→ risk of leakage via reverseengineering — increases Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concept_edges[6] |
| note | training process is exposed to multiple parties —causes→ risk of leakage via reverseengineering — The paper directly says multi-party exposure increases reverse-engineering leakage risk. Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concept_edges[6] |
| audit | secure encryption methods —causes→ data leakage — Audit 2026-09-28: DOUBTFUL - read against its full span and kept, but the span names designing secure encryption methods to prevent leakage as a CHALLENGE; causes is loose. Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concept_edges[7] |
| effect direction | secure encryption methods —causes→ data leakage — reduces Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concept_edges[7] |
| note | secure encryption methods —causes→ data leakage — Secure encryption is named as an architecture measure to prevent data leakage. Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concept_edges[7] |
| audit | FL —causes→ accuracy and robustness of healthcare AI — Audit 2026-09-28: DOUBTFUL - read against its full span and kept, but the span says FL HAS THE POTENTIAL to increase accuracy and robustness; hedged. Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concept_edges[8] |
| effect direction | FL —causes→ accuracy and robustness of healthcare AI — increases Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concept_edges[8] |
| note | FL —causes→ accuracy and robustness of healthcare AI — Weight 0.7 preserves the paper's 'has the potential to increase' wording. Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concept_edges[8] |
| weight | FL —causes→ accuracy and robustness of healthcare AI — 0.7 Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concept_edges[8] |
| audit | FL —causes→ costs — Audit 2026-09-28: DOUBTFUL - read against its span and kept, but the span shown is about paradigm shifts in healthcare payment; the cost claim is not visible. Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concept_edges[9] |
| effect direction | FL —causes→ costs — reduces Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concept_edges[9] |
| note | FL —causes→ costs — The projected cost reduction shares the paper's potential-language hedge. Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concept_edges[9] |
| weight | FL —causes→ costs — 0.7 Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concept_edges[9] |
| audit | FL —causes→ patient outcomes — Audit 2026-09-28: DOUBTFUL - read against its full span and kept, but the span says FL HAS THE POTENTIAL to improve patient outcomes; hedged. Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concept_edges[10] |
| effect direction | FL —causes→ patient outcomes — increases Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concept_edges[10] |
| note | FL —causes→ patient outcomes — The projected improvement in patient outcomes is weighted 0.7 for the paper's modal framing. Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concept_edges[10] |
| weight | FL —causes→ patient outcomes — 0.7 Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concept_edges[10] |
| evidence status note | concept — THE QUOTE CONTINUES A SENTENCE THAT THIS COPY CUTS IN TWO - a figure caption (the FL compute plans panels) stands between it and the words it follows, which are not adjacent to it in the extraction. Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[45] |
| evidence status note | concept — THE QUOTE CONTINUES A SENTENCE THAT THIS COPY CUTS IN TWO - a figure caption (the FL compute plans panels) stands between it and the words it follows, which are not adjacent to it in the extraction. Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[46] |
| evidence status note | concept — THE QUOTE CONTINUES A SENTENCE THAT THIS COPY CUTS IN TWO - a figure caption (the FL compute plans panels) stands between it and the words it follows, which are not adjacent to it in the extraction. Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[47] |
| no type for this | FL → Federated learning — FL expands to Federated learning. Acronym expansion is absent from the eight edge types. Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | unhoused_relations[0] |
| no type for this | FL → ML from non-co-located data — FL enables machine learning from non-co-located data. Enablement is absent from the eight edge types and is not safely reducible to causation here. Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | unhoused_relations[1] |
| no type for this | Non-trusted → security strategies — Non-trusted FL collaborations require security strategies. Prerequisite-for is absent from the eight edge types. Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | unhoused_relations[2] |
| no type for this | traceability and accountability processes → execution integrity — Traceability and accountability in non-trusted federations require execution integrity. Prerequisite-for is absent from the eight edge types. Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | unhoused_relations[3] |
| no type for this | “honest broker” system → trusted third party — An honest-broker federation is administered through a trusted third party acting as intermediary. Administered-through or mediated-by is absent from the eight edge types. Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | unhoused_relations[4] |
| limitation | FL is stated to still require rigorous technical consideration to avoid compromising safety or patient privacy - the paradigm is offered as promising rather than solved Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | limitations[0] |
| limitation | The authors say plainly that not all technical questions have been answered and expect FL to remain an open research area for a decade Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | limitations[1] |
| limitation | The performance claim is attributed to other work rather than demonstrated here - this is a perspective paper citing comparisons, not running them Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | limitations[2] |
| limitation | Public medical data carries varying license restrictions that sometimes limit its exploitation, which bounds what any centralised alternative could have used anyway Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | limitations[3] |
| concept | Federated learning Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[0] |
| concept | data governance Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[1] |
| concept | consensus model Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[2] |
| concept | anonymisation Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[3] |
| concept | reidentification Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[4] |
| concept | privacy Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[5] |
| concept | precision medicine Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[6] |
| concept | license restrictions Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[7] |
| concept | egress Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[8] |
| concept | business value Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[9] |
| concept | firewalls Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[10] |
| concept | Data-driven machine learning (ML) — not defined in this paper's own words; medical-model development from large healthcare data volumes Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[11] |
| concept | accurate and robust statistical models — not defined in this paper's own words; target models built from medical data Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[12] |
| concept | data silos — not defined in this paper's own words; institutional separation that limits exploitation of medical data Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[13] |
| concept | privacy concerns — not defined in this paper's own words; second barrier restricting access to medical data Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[14] |
| concept | without access to sufficient data — not defined in this paper's own words; data-access condition that blocks ML deployment Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[15] |
| concept | full potential — not defined in this paper's own words; capability ML cannot reach without sufficient data Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[16] |
| concept | transition from research to clinical practice — not defined in this paper's own words; deployment transition blocked by insufficient data Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[17] |
| concept | Centralised Data Lake — not defined in this paper's own words; pooled-data alternative contrasted with federated workflows Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[18] |
| concept | FL aggregation server — not defined in this paper's own words; federated workflow coordinated by a central aggregation server Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[19] |
| concept | federation of training nodes — not defined in this paper's own words; participants receiving and returning partially trained models Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[20] |
| concept | global model — not defined in this paper's own words; model distributed to training nodes Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[21] |
| concept | partially trained models — not defined in this paper's own words; local model states resubmitted for aggregation Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[22] |
| concept | central server — not defined in this paper's own words; coordinator that aggregates participant models Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[23] |
| concept | FL peer to peer — not defined in this paper's own words; alternative workflow with local exchange and aggregation Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[24] |
| concept | Centralised training — not defined in this paper's own words; non-federated workflow donating data to a central lake Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[25] |
| concept | local, independent training — not defined in this paper's own words; training performed after extracting data from a central lake Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[26] |
| concept | ML from non-co-located data — not defined in this paper's own words; core federated promise for privacy and governance Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[27] |
| concept | data controller — not defined in this paper's own words; party defining governance, privacy, and access policy Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[28] |
| concept | governance processes — not defined in this paper's own words; locally defined rules for federated participation Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[29] |
| concept | privacy policies — not defined in this paper's own words; locally defined privacy constraints Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[30] |
| concept | data access — not defined in this paper's own words; controller capability that can be revoked Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[31] |
| concept | validation phase — not defined in this paper's own words; phase covered by local access control Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[32] |
| concept | large-scale, ininstitutional validation — not defined in this paper's own words; validation opportunity enabled without pooling data Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[33] |
| concept | rare diseases — not defined in this paper's own words; research setting where each institution has too little data Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[34] |
| concept | high-dimensional, storage-intense medical data — not defined in this paper's own words; data kept at local institutions rather than duplicated Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[35] |
| concept | centralised pool — not defined in this paper's own words; storage-intensive pooling arrangement avoided by FL Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[36] |
| concept | growing global data set — not defined in this paper's own words; distributed data collection with which the model can scale Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[37] |
| concept | data storage requirements — not defined in this paper's own words; resource demand not increased disproportionately by model-to-data transfer Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[38] |
| concept | FL topologies — not defined in this paper's own words; communication architectures of a federation Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[39] |
| concept | Centralised — not defined in this paper's own words; topology coordinated by an aggregation server Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[40] |
| concept | Decentralised — not defined in this paper's own words; topology with peer connections and local parallel aggregation Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[41] |
| concept | Hierarchical — not defined in this paper's own words; topology composed from mixed sub-federations Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[42] |
| concept | FL compute plans — not defined in this paper's own words; model trajectories across several partners Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[43] |
| concept | Sequential training/cyclic transfer learning — not defined in this paper's own words; compute plan moving a model sequentially Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[44] |
| concept | memorise information — not defined in this paper's own words; privacy risk carried by trained models under some conditions Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[45] |
| concept | differential privacy — not defined in this paper's own words; mechanism proposed to enhance federated privacy Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[46] |
| concept | learning from encrypted data — not defined in this paper's own words; second privacy-enhancing mechanism proposed for FL Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[47] |
| concept | volume-based — not defined in this paper's own words; fee-for-service healthcare model Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[48] |
| concept | fee-for-service-based — not defined in this paper's own words; expansion of volume-based healthcare Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[49] |
| concept | value-based healthcare — not defined in this paper's own words; outcomes-focused model linked to precision medicine Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[50] |
| concept | focused treatment — not defined in this paper's own words; route to earlier outcomes and lower cost Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[51] |
| concept | accuracy and robustness of healthcare AI — not defined in this paper's own words; model qualities FL may increase Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[52] |
| concept | costs — not defined in this paper's own words; resource burden FL may reduce Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[53] |
| concept | patient outcomes — not defined in this paper's own words; clinical result FL may improve Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[54] |
| concept | FL — not defined in this paper's own words; abbreviation used for federated learning Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[55] |
| concept | learning paradigm — not defined in this paper's own words; class assigned to federated learning Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[56] |
| concept | multiple parties — not defined in this paper's own words; collaborators training without centralising their data Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[57] |
| concept | global loss function — not defined in this paper's own words; weighted combination of local losses Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[58] |
| concept | weighted combination of K local losses — not defined in this paper's own words; general mathematical formulation of global loss Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[59] |
| concept | private data — not defined in this paper's own words; party-resident input never shared among collaborators Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[60] |
| concept | data heterogeneity — not defined in this paper's own words; cross-site diversity creating federated optimisation challenges Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[61] |
| concept | inhomogeneous data distribution — not defined in this paper's own words; non-IID distribution challenging FL algorithms Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[62] |
| concept | independently and identically distributed (IID) data — not defined in this paper's own words; assumption violated by many medical federations Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[63] |
| concept | FedAvg — not defined in this paper's own words; aggregation strategy prone to failure under inhomogeneous distributions Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[64] |
| concept | FedProx — not defined in this paper's own words; strategy addressing heterogeneous federated data Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[65] |
| concept | part-data-sharing strategy — not defined in this paper's own words; second strategy addressing heterogeneous data Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[66] |
| concept | FL with domainadaptation — not defined in this paper's own words; third strategy addressing heterogeneous data Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[67] |
| concept | global optimal solution — not defined in this paper's own words; federation-wide optimum that may not serve each participant Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[68] |
| concept | optimal for an individual local participant — not defined in this paper's own words; local optimality contrasted with the global solution Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[69] |
| concept | model training optimality — not defined in this paper's own words; criterion participants should agree before training Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[70] |
| concept | Privacy-preserving techniques for FL — not defined in this paper's own words; protections exceeding contemporary commercial ML models Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[71] |
| concept | levels of protection — not defined in this paper's own words; privacy benefit offered by protection techniques Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[72] |
| concept | accuracy of the final model — not defined in this paper's own words; performance dimension privacy techniques may affect Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[73] |
| concept | future techniques — not defined in this paper's own words; future capabilities that could compromise a low-risk model Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[74] |
| concept | ancillary data — not defined in this paper's own words; additional data that could compromise a low-risk model Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[75] |
| concept | low-risk — not defined in this paper's own words; prior model risk assessment vulnerable to later compromise Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[76] |
| concept | Non-trusted — not defined in this paper's own words; large-scale collaboration without enforceable trust agreements Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[77] |
| concept | security strategies — not defined in this paper's own words; countermeasures required in non-trusted federations Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[78] |
| concept | advanced encryption of model submissions — not defined in this paper's own words; non-trusted federation security strategy Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[79] |
| concept | secure authentication of all parties — not defined in this paper's own words; identity security strategy Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[80] |
| concept | traceability of actions — not defined in this paper's own words; accountability strategy for non-trusted federation Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[81] |
| concept | verification systems — not defined in this paper's own words; integrity strategy for non-trusted federation Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[82] |
| concept | execution integrity — not defined in this paper's own words; required property of accountable non-trusted processing Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[83] |
| concept | model confidentiality — not defined in this paper's own words; protected model property Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[84] |
| concept | protections against adversarial attacks — not defined in this paper's own words; security strategy for malicious clients Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[85] |
| concept | traditional training — not defined in this paper's own words; training mode contrasted with multi-party exposure Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[86] |
| concept | training process is exposed to multiple parties — not defined in this paper's own words; federated exposure increasing reverse-engineering leakage risk Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[87] |
| concept | risk of leakage via reverseengineering — not defined in this paper's own words; privacy risk increased by observing federated model changes Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[88] |
| concept | multiparty computations — not defined in this paper's own words; distributed computations requiring traceable configurations Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[89] |
| concept | training configurations — not defined in this paper's own words; configuration state that must be traced Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[90] |
| concept | hyperparameter tuning — not defined in this paper's own words; training activity requiring traceability Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[91] |
| concept | contribution from each participant — not defined in this paper's own words; measured input used for compensation and revenue models Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[92] |
| concept | computational resources consumed — not defined in this paper's own words; one participant-contribution measure Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[93] |
| concept | quality of the data used for local training — not defined in this paper's own words; second participant-contribution measure Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[94] |
| concept | compensation — not defined in this paper's own words; allocation determined from participant contribution Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[95] |
| concept | revenue model — not defined in this paper's own words; federation economics based on participant measurements Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[96] |
| concept | secure intra-node viewing facility — not defined in this paper's own words; controlled way to inspect local data Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[97] |
| concept | explainability and interpretability of the global model — not defined in this paper's own words; goal of controlled data-viewing support Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[98] |
| concept | redundant nodes — not defined in this paper's own words; architecture measure for communication data integrity Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[99] |
| concept | secure encryption methods — not defined in this paper's own words; architecture measure preventing data leakage Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[100] |
| concept | data leakage — not defined in this paper's own words; risk targeted by secure encryption Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[101] |
| concept | node schedulers — not defined in this paper's own words; resource-allocation mechanism reducing idle time Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[102] |
| concept | “honest broker” system — not defined in this paper's own words; federation administered through a trusted intermediary Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[103] |
| concept | trusted third party — not defined in this paper's own words; intermediary controlling an honest-broker system Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[104] |
| concept | peer-topeer system — not defined in this paper's own words; federation in which sites interact directly Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[105] |
| concept | gatekeeper function — not defined in this paper's own words; central control absent from peer-to-peer administration Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[106] |
| concept | trustlessbased architecture — not defined in this paper's own words; architecture cryptographically constraining the operator Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[107] |
| concept | secure protocol — not defined in this paper's own words; cryptographic means of locking the operator into honest behavior Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[108] |
| concept | computational overheads — not defined in this paper's own words; cost introduced by trustless protocols Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[109] |
When we have changed our mind. Some entries in the table begin with a word in capitals and the reason for it. EVIDENCE CORRECTED — the finding still stands, but we had quoted the wrong sentence from the paper, and the quotation has been replaced with the right one. RETRACTED — the entry itself was wrong; it stays on the page, marked, with the reason, rather than being deleted. WITHDRAWN AND REPLACED — the entry was wrong, and a corrected version of it appears in the entries below. Each mark carries its own reason, in brackets, because the reasons differ from one entry to the next. An entry we have WITHDRAWN is removed from this page and kept in the record, with its reason. A line reading Audit and a date is a later re-reading of the entry against its whole sentence: DOUBTFUL means it was kept, and the reason says what is loose about it. Quotation marks inside the tagger's sentence. Where a tag is the tagger's own sentence, a phrase in quotation marks inside it is borrowed from the paper — the tagger's wording around it is not. The paper's own sentence is the evidence quote for that row; on this page it is withheld with the article, and the article is at its publisher. Nothing inside quotation marks anywhere on this page is ours: if it is quoted, it is the paper's. Whose words a concept carries. A concept entry records a term the paper itself uses, so the term, the sense given for it, and its evidence are the paper's own words, quoted where the paper fixes them. This is the opposite of a sense entry in the tag table, which is the tagger's sentence about how the paper uses a term. Where a concept entry could not be grounded in the paper's words, the entry says so on its own row.
Tag layer © the Meta-Tagging Project, CC BY 4.0. The paper text is © its authors/publisher
and is linked, not reproduced. Excerpts shown are short quotations for identification and verification.
· back to index
· concepts · authors
· עברית
© 2026 Shir Sivroni · ORCID 0009-0000-2597-2824