(cid:NN). The PDF this text was extracted from encodes 5 characters in a way that names no letter, so the extractor could not say what they were. The quotes keep them because a quote is reproduced exactly or it proves nothing — the marks are the source’s, not the tagging’s.phenomenon · claim-type · cross-domain-transfer · mechanism · claim-relation · replication-status · limitation — grounded in the full text| phenomenon | job-shop scheduling— “The job shop scheduling problem (JSSP) is, a type of scheduling problem that aims to determine the optimal sequential assignments of machines to multiple jobs consisting of series of operations while preserving the problem constraints (processing precedence and machine-sharing).” | |
| claim-type | method— “We propose a framework to learn to schedule a job-shop problem (JSSP) using a graph neural network (GNN) and reinforcement learning (RL).” | |
| method — evidence | — “In this paper, we propose a framework to construct the scheduling policy for JSSPs using GNN and RL.” | |
| transfer | from machine learning (graph representation learning): The size-agnostic transferability of graph neural networks, demonstrated in robot control, wind power estimation and weather prediction, is imported into production scheduling to make a scheduler that works on unseen shop sizes— “We utilize such transferability of GNNs to accomplish the transferable property in our JSSP scheduler.” | transfers[0] |
| transfer | from deep reinforcement learning: Proximal Policy Optimization, a policy-gradient algorithm from the RL literature, is adopted to train the representation and dispatching modules of a job-shop scheduler end to end— “We use proximal policy optimization (PPO) [37] algorithm, a variant of policy-based RL, to train the GNN based state representation module and the parameterized decision-making policy jointly.” | transfers[1] |
| mechanism | Because the embedding layer works through pairwise interaction functions that are shared across all node pairs rather than tied to a fixed graph size, the same trained network can be run on graphs with different node counts and edge patterns, which is what carries the policy to unseen instance sizes— “Due to the pair-wise interaction module, a trained GNN can be applied to the news graphs whose structures (the number of nodes in graph and the edge connectivities) are different from the graphs used for training.” | mechanisms[0] |
| mechanism | Training on a graph representation forces the policy to key on the structural regularities common to all job shops instead of memorising one instance, which is why it transfers with no further training— “The scheduling policy learns to consider the inherent structure of the JSSP, rather than learning the characteristics of a specific JSSP instance.” | mechanisms[1] |
| mechanism | The learned policy beats fixed rules because it re-weights which state features matter at each decision, whereas a dispatching rule always consults the same predetermined features— “The key advantages of the RL-based approach over the PDR and CDR are that it learns how to adaptively utilize various state features in determining the optimal scheduling actions, while PDRs and CDRs constantly use the predetermined same state features in making the next scheduling actions.” | mechanisms[2] |
| mechanism | The dense waiting-job reward is used instead of makespan because a reward delivered only at episode end leaves the algorithm unable to assign credit to individual dispatching actions— “On the other hand, the makespan reward will be realized once only when the given instance is finished; as a result, the RL algorithm has difficulty in assigning credit to the actions in the episode during training.” | mechanisms[3] |
| builds-on | Proximal Policy Optimization (Schulman et al., cited as [37]/[38])— “We use proximal policy optimization (PPO) [37] algorithm, a variant of policy-based RL, to train the GNN based state representation module and the parameterized decision-making policy jointly.” | claim_relations[0] |
| extends | MARL [10] and MDQN [24] RL-based JSSP schedulers— “To mimic the experimental procedures reported in [10] and [24], we trained the three GNN schedulers (GNN1, GNN2, and GNN3) with the training JSSP instances with sizes 5×15, 10×10 and 15×20, respectively.” | claim_relations[1] |
| sense | generalization - AS THE ABSTRACT CLAIMS IT: a superb capability, asserted without a scope— “We empirically demonstrate that the GNN scheduler, due to its superb generalization capability, outperforms practically favored dispatching rules and RL-based schedulers on various benchmark JSSP.” | senses[0] |
| sense | generalization - AS THE CONCLUSION SCOPES IT: trained on only small JSSPs and producing high-quality schedules for large ones, with the PDR comparison run on the training distribution— “Overall, the results demonstrate that our framework can produce a generalizable scheduling policy so that the GNN scheduler trained on only small JSSPs can generate high-quality schedules for large-sized JSSPs without additional training.” | senses[1] |
| sense | scheduling quality - operationalised entirely as makespan: the maximum completion time among jobs, which is the single quantity the reward function is built to minimise— “We aim to learn a scheduling policy that minimizes the makespan Cmax. Cmax is the maximum completion time among jobs, i.e., the total length of the schedule” | senses[2] |
| control | the 6x6 instances were dropped from the performance evaluation because only one benchmark problem exists at that size — held — rules out that a mean over problem sizes is dominated by a size represented by a single instance— “We excluded the 6×6 (6 machines and 6 jobs) instances from the performance evaluation because the benchmark JSSP with that size is only FT06.” | controls[0] |
| audit | job shop scheduling problem —operationalised-as→ MDP — Audit 2026-09-28: DOUBTFUL - read against its full span and kept, but the span writes 'JSSP'; endpoint abbreviated.— “We formulate a scheduling process of a JSSP instance as an MDP.” | concept_edges[0] |
| outcome (stated by the tagger, no quotation) | control — held | controls[0] |
| for which use | limitation — the practical scalability claim, not the solution quality: THE COMPUTATIONAL TIME TO COMPUTE THE NODE EMBEDDINGS RAPIDLY INCREASES with instance size— “Although the proposed GNN scheduler showed reliable performance on our testing set and benchmark JSSP instances, the computational time to compute the node embeddings rapidly increases as the size of JSSP increases.” | limitations[0] |
| for which use | limitation — any shop-floor deployment reading: THE CURRENT FORMULATION DOES NOT CONSIDER MORE PRACTICAL CONSTRAINTS SUCH AS BATCH PROCESSES AND WAITING TIME REGULATIONS AND FLEXIBLE SHOP SETUP— “Lastly, the current formulation does not consider more practical constraints such as batch processes and waiting time regulations and flexible shop setup.” | limitations[1] |
| for which use | limitation — any claim of applicability to real job shops: THE DISTRIBUTION P0(g0) USED TO GENERATE THE TRAINING AND TEST JSSP INSTANCES IS NOT GENERAL ENOUGH TO COVER REAL-WORLD JSSPs— “In a practical sense, the distribution P0(g0) used to generate the training and test JSSP instances is not general enough to cover real-world JSSPs.” | limitations[2] |
| for which use | limitation — the comparison against prior RL schedulers: ONLY A LIMITED GENERALIZATION TEST COULD BE CONDUCTED FOR THE CASE WHERE THE CONJUNCTIVE CONSTRAINTS AND PROCESSING TIMES ARE DIFFERENT FROM THE TRAINING AND TEST DATA SETS— “Therefore, only a limited generalization test could be conducted for the case where the conjunctive constraints and processing times are different from the training and test data sets.” | limitations[3] |
| for which use | limitation — any deployment where scheduling latency matters: THE PROPOSED GNN-SCHEDULER REQUIRES MORE COMPUTATIONAL TIME THAN PDRs— “As shown in Figure 8, the proposed GNN-scheduler requires more computational time than PDRs.” | limitations[4] |
| for which use | limitation — the claim that the GNN scheduler beats dispatching rules and other RL schedulers in general: the PDR comparison is on instances from the training distribution, and the RL comparison carries a stated restriction on testing instances— “The proposed GNN scheduler outperformed all PDRs when testing on JSSP instances generated by the distribution used for generating training instances. The proposed GNN scheduler showed robust performances on various sets of benchmark instances in contrast to PDRs, whose solution qualities fluctuated depending on the properties of the given… | limitations[5] |
| for which use | limitation — applying this scheduler to large shops: the authors state embedding time grows in proportion to n2 + nm in the numbers of jobs and machines— “the computational time to compute the node embeddings rapidly increases as the size of JSSP increases”— “the computational time to compute the node embeddings is also proportional to n2 + nm.” | limitations[6] |
| for which use | limitation — the optimality gaps reported for the validation set: the reference minimum makespan is computed by the Google OR tool rather than supplied independently— “We prepare the validation instances generated through p0(G0). The optimal solutions (minimum makespan) of the validation instances are computed through the Google OR tool” | limitations[7] |
| for which use | limitation — the trained policy as a fixed artefact: training was terminated when the validation-set makespan converged, so the model is defined by that stopping rule— “We terminate the training procedure when the makespan of the validation set converges.” | limitations[8] |
| for which use | limitation — the reported architecture as tuned: deeper and thicker networks were tested and are reported not to have improved performance, without the comparison being shown— “We tested deeper and thicker ReLU networks while finding the best hyperparameters. The deep and thick ReLU networks than the specified hyperparameter setup did not improve scheduling performance.” | limitations[9] |
| replication | untested | |
| limitation | Computation time for node embeddings grows rapidly (quadratically in the number of jobs) with instance size.— “Although the proposed GNN scheduler showed reliable performance on our testing set and benchmark JSSP instances, the computational time to compute the node embeddings rapidly increases as the size of JSSP increases.” | limitations[0] |
| limitation | The MDP formulation omits practical shop-floor constraints such as batching, waiting-time regulations and flexible setups.— “Lastly, the current formulation does not consider more practical constraints such as batch processes and waiting time regulations and flexible shop setup.” | limitations[1] |
| limitation | The distribution used to generate training and test instances does not cover real-world job shops.— “In a practical sense, the distribution P0(g0) used to generate the training and test JSSP instances is not general enough to cover real-world JSSPs.” | limitations[2] |
| limitation | The comparison with prior RL schedulers could only test limited generalisation, because those baselines require train and test instances of the same size.— “Therefore, only a limited generalization test could be conducted for the case where the conjunctive constraints and processing times are different from the training and test data sets.” | limitations[3] |
| limitation | The learned scheduler is slower at inference than simple priority dispatching rules.— “As shown in Figure 8, the proposed GNN-scheduler requires more computational time than PDRs.” | limitations[4] |
| limitation | The headline comparison is scoped to instances drawn from the training distribution, and the comparison against the other learned schedulers is scoped by a restriction on testing instances. | limitations[5] |
| limitation | The method's cost grows with problem size by the authors' own analysis, which is the quantity a scheduling practitioner would care about. | limitations[6] |
| limitation | The reference against which solution quality is judged is another solver's output, not a proven optimum supplied independently. | limitations[7] |
| limitation | Training stops on a convergence criterion measured on the validation set, so the reported model is the one that stopped improving on that set. | limitations[8] |
| limitation | The architecture search behind the reported hyperparameters is reported only as a negative result, with no numbers attached. 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[9] |
finding · quantity · variable · definition · key-term — grounded in the full text| finding | On test instances the trained GNN scheduler reached 14.7% average relative error. 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. | content_tags.findings[0] |
| finding | On published benchmark sets the GNN scheduler gave average relative errors from 12.6% to 20.9%. 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. | content_tags.findings[1] |
| finding | A scheduler trained only on small randomly generated instances transfers to much larger instances, since the gap between the generically trained GNN0 and the size-matched GNN1-3 was small. 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. | content_tags.findings[2] |
| finding | The GNN scheduler reliably produced high-quality solutions across instance 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. | content_tags.findings[3] |
| finding | Computation time grows roughly with the square of the number of jobs (quadratic fit R2 = 0.998). 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. | content_tags.findings[4] |
| finding | Stacking more than three embedding layers bought no measurable scheduling improvement while costing extra computation, so the message-passing depth was fixed at three. 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. | content_tags.findings[5] |
| finding | The GNN scheduler needs more computation time than the priority dispatching rules. 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. | content_tags.findings[6] |
| finding | Fewer than 20 episodes per parameter update left the state-value estimate with large variance, while more than 20 added no performance, fixing the batch size at 20 episodes. 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. | content_tags.findings[7] |
| quantity | 14.7% — average relative makespan error of the trained scheduler on the 100 in-distribution test instances 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. | content_tags.quantities[0] |
| quantity | over 20% — average relative error of every dispatching rule except MOR on the same test set 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. | content_tags.quantities[1] |
| quantity | 12.6% to 20.9% — range of the scheduler's average relative error across the published benchmark families 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. | content_tags.quantities[2] |
| quantity | 4.4% and 2.7% — the two cases where the scheduler was second-best by instance size (20x30 and 20x50), and by how much it trailed the best rule 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. | content_tags.quantities[3] |
| quantity | R2 = 0.9984 — fit quality of the quadratic model relating runtime to the number of jobs 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. | content_tags.quantities[4] |
| quantity | R2 = 0.9815 — fit quality of the linear model relating runtime to the number of machines 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. | content_tags.quantities[5] |
| quantity | K = 3 — number of stacked embedding layers (hops of message passing) used in the GNN 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. | content_tags.quantities[6] |
| quantity | 100 — number of held-out test instances sampled from the training distribution 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. | content_tags.quantities[7] |
| quantity | 20 machines and 100 jobs — size of the largest instance the scheduler was tested on 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. | content_tags.quantities[8] |
| quantity | 20 — episodes collected per PPO parameter update 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. | content_tags.quantities[9] |
| variable | number of jobs (independent) 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. | content_tags.variables[0] |
| variable | number of machines (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. | content_tags.variables[1] |
| variable | relative scheduling error (dependent) 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. | content_tags.variables[3] |
| variable | computation time (dependent) 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. | content_tags.variables[4] |
| variable | additional training (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. | content_tags.variables[5] |
| variable | discount factor (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. | content_tags.variables[7] |
| variable | learning rate (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. | content_tags.variables[8] |
| evidence status note | variable — RULED, NOT PREVIOUSLY ADJUDICATED. This quote carries an unmapped font glyph and was not covered by the 2026-08-25 ruling, which reached only rows flagged with glyph damage. The same reasoning applies and the same decision follows: THE QUOTE IS NOT EDITED. The unmapped characters are the ones carrying the claim, and no code-to-character table is possible - in one sentence of this corpus the same code is '=' once and 'tau' twice. Repair needs the source PDF and is a re-staging job, costed and refused (0 of 114 appendable; a per-record re-stage fixes 21 and breaks 223). Three sibling rows were found ALREADY ruled under this same pass and are untouched. | content_tags.variables[8] |
| variable | GAE parameter (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. | content_tags.variables[9] |
| evidence status note | variable — RULED, NOT PREVIOUSLY ADJUDICATED. This quote carries an unmapped font glyph and was not covered by the 2026-08-25 ruling, which reached only rows flagged with glyph damage. The same reasoning applies and the same decision follows: THE QUOTE IS NOT EDITED. The unmapped characters are the ones carrying the claim, and no code-to-character table is possible - in one sentence of this corpus the same code is '=' once and 'tau' twice. Repair needs the source PDF and is a re-staging job, costed and refused (0 of 114 appendable; a per-record re-stage fixes 21 and breaks 223). Three sibling rows were found ALREADY ruled under this same pass and are untouched. | content_tags.variables[9] |
| variable | clipping parameter (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. | content_tags.variables[10] |
| evidence status note | variable — RULED, NOT PREVIOUSLY ADJUDICATED. This quote carries an unmapped font glyph and was not covered by the 2026-08-25 ruling, which reached only rows flagged with glyph damage. The same reasoning applies and the same decision follows: THE QUOTE IS NOT EDITED. The unmapped characters are the ones carrying the claim, and no code-to-character table is possible - in one sentence of this corpus the same code is '=' once and 'tau' twice. Repair needs the source PDF and is a re-staging job, costed and refused (0 of 114 appendable; a per-record re-stage fixes 21 and breaks 223). Three sibling rows were found ALREADY ruled under this same pass and are untouched. | content_tags.variables[10] |
| variable | value function coefficient (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. | content_tags.variables[11] |
| evidence status note | variable — RULED, NOT PREVIOUSLY ADJUDICATED. This quote carries an unmapped font glyph and was not covered by the 2026-08-25 ruling, which reached only rows flagged with glyph damage. The same reasoning applies and the same decision follows: THE QUOTE IS NOT EDITED. The unmapped characters are the ones carrying the claim, and no code-to-character table is possible - in one sentence of this corpus the same code is '=' once and 'tau' twice. Repair needs the source PDF and is a re-staging job, costed and refused (0 of 114 appendable; a per-record re-stage fixes 21 and breaks 223). Three sibling rows were found ALREADY ruled under this same pass and are untouched. | content_tags.variables[11] |
| variable | entropy bonus coefficient (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. | content_tags.variables[12] |
| evidence status note | variable — RULED, NOT PREVIOUSLY ADJUDICATED. This quote carries an unmapped font glyph and was not covered by the 2026-08-25 ruling, which reached only rows flagged with glyph damage. The same reasoning applies and the same decision follows: THE QUOTE IS NOT EDITED. The unmapped characters are the ones carrying the claim, and no code-to-character table is possible - in one sentence of this corpus the same code is '=' once and 'tau' twice. Repair needs the source PDF and is a re-staging job, costed and refused (0 of 114 appendable; a per-record re-stage fixes 21 and breaks 223). Three sibling rows were found ALREADY ruled under this same pass and are untouched. | content_tags.variables[12] |
| definition | job shop scheduling problem (JSSP) 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. | content_tags.definitions[0] |
| definition | priority dispatching rule 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. | content_tags.definitions[1] |
| definition | graph state for the JSSP 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. | content_tags.definitions[2] |
| definition | non-trivial state 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. | content_tags.definitions[3] |
| definition | primitive 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. | content_tags.definitions[4] |
| definition | graph neural network (GNN) 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. | content_tags.definitions[5] |
| definition | makespan 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. | content_tags.definitions[6] |
| concept | job shop scheduling problem 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 | MDP 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 | makespan 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 | priority dispatching rules 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] |
| regression (polynomial curve fitting) | as written: the fitted second-order polynomial curve 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. | content_tags.statistics[0] |
| coefficient of determination R2 | as written: R2 = 0.9815 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. | content_tags.statistics[1] |
| standard deviation | as written: mean and standard deviation 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. | content_tags.statistics[2] |
| error bars (one standard deviation) | as written: one standard deviation of errors 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. | content_tags.statistics[3] |
| mean | as written: average relative error 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. | content_tags.statistics[4] |
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