{"id":"0179e71f4d5d7f5362533c8adc9e0aba7d3f62136af881944c16233ab1232488","title":"Pavlovic, Paun and Poesio argue human soft-labels matter more for calibration than accuracy","argument":"Maja Pavlovic, Silviu Paun and Massimo Poesio argue that human soft-labels—labels that retain uncertainty rather than giving only one answer—may matter less for improving accuracy than for improving calibration: whether a model’s confidence reflects its uncertainty. Their paper reports a controlled audit on MNIST and a synthetic variant, designed to separate the effect of soft-labels from mode shifts, or changes in which label is treated as the main answer. The authors say human soft-labels improved accuracy, but had a larger value as a regularizer, helping calibration on difficult samples and stable convergence across training runs. They also report that dataset cartography showed models trained on human soft-labels mirrored human uncertainty, while models trained on synthetic labels did not align with humans. The distinction matters because earlier studies, they argue, conflated benefits of soft-labels with correction of mislabeled data. The authors present their work as a diagnostic testbed for human-AI uncertainty alignment, not evidence here of performance across other datasets or settings.","topicIds":["intelligence-and-discovery"],"domain":"science","claimType":"attributed reading","limits":"The supplied evidence is the paper’s abstract and publisher page; it does not provide detailed methods, measurements, or results beyond the summary presented there. The reported audit covers MNIST and a synthetic variant, so the supplied evidence does not establish whether the findings extend to other datasets or settings.","speaker":{"name":"Pavlovic, Maja, Paun, Silviu, Poesio, Massimo","roleAtSource":"Article author"},"source":{"title":"An Assessment of Human vs. Model Uncertainty in Soft-Label Learning and Calibration","publisher":"arXiv cs.AI","format":"Article","url":"https://arxiv.org/abs/2605.18648","publishedAt":"2026-05-18","dateBasis":"Date basis not provided."},"evidence":[{"text":"The authors argue earlier work conflated benefits of soft-labels with correction of mislabeled data, obscuring the effects of soft-label supervision itself.","excerpt":"prior studies conflate these benefits with the implicit correction of mislabeled data (mode shifts), obscuring true effects of soft-labels.","url":"https://arxiv.org/abs/2605.18648","relationship":"Source-stated"},{"text":"In an audit using MNIST and a synthetic variant, the authors report that human soft-labels improved accuracy.","excerpt":"across MNIST and a synthetic variant, re-annotating subsets to extract human uncertainty. By decoupling soft-label supervision from underlying label mode shifts, we show that while human soft-labels do provide accuracy gains","url":"https://arxiv.org/abs/2605.18648","relationship":"Source-stated"},{"text":"The authors say human soft-labels’ larger value was improved calibration on difficult samples and more stable convergence across training runs.","excerpt":"their larger value lies in acting as a regularizer that improves model calibration on difficult samples and promotes stable convergence across training runs.","url":"https://arxiv.org/abs/2605.18648","relationship":"Source-stated"},{"text":"The authors report that models trained on human soft-labels mirrored human uncertainty, unlike models trained on synthetic labels.","excerpt":"models trained on human soft-labels mirror human uncertainty, whereas those trained on synthetic labels fail to align with humans.","url":"https://arxiv.org/abs/2605.18648","relationship":"Source-stated"},{"text":"The authors describe the work as a diagnostic testbed for human-AI uncertainty alignment.","excerpt":"Broadly, this work provides a diagnostic testbed for human-AI uncertainty alignment.","url":"https://arxiv.org/abs/2605.18648","relationship":"Source-stated"}],"sourceMedia":{"status":"not_found","sourceUrl":"https://arxiv.org/abs/2605.18648","checkedAt":"2026-11-08T21:04:08.245Z","items":[],"note":"No suitable media was found by the supported page extraction."},"assumptions":[],"openQuestion":{"attribution":"Open question","text":"Do human soft-labels produce the reported calibration and uncertainty-alignment patterns beyond MNIST and the synthetic variant?"},"whatWouldChangeView":{"attribution":"Change condition not recorded","text":""},"editorialComparisons":[],"live":{"title":"Pavlovic, Paun and Poesio argue human soft-labels matter more for calibration than accuracy","summary":"Maja Pavlovic, Silviu Paun and Massimo Poesio argue that human soft-labels—labels that retain uncertainty rather than giving only one answer—may matter less for improving accuracy than for improving calibration: whether a model’s confidence reflects its uncertainty. Their paper reports a controlled audit on MNIST and a synthetic variant, designed to separate the effect of soft-labels from mode shifts, or changes in which label is treated as the main answer. The authors say human soft-labels improved accuracy, but had a larger value as a regularizer, helping calibration on difficult samples and stable convergence across training runs. They also report that dataset cartography showed models trained on human soft-labels mirrored human uncertainty, while models trained on synthetic labels did not align with humans. The distinction matters because earlier studies, they argue, conflated benefits of soft-labels with correction of mislabeled data. The authors present their work as a diagnostic testbed for human-AI uncertainty alignment, not evidence here of performance across other datasets or settings.","question":"Do human soft-labels produce the reported calibration and uncertainty-alignment patterns beyond MNIST and the synthetic variant?","claims":[{"text":"The authors argue earlier work conflated benefits of soft-labels with correction of mislabeled data, obscuring the effects of soft-label supervision itself.","excerpt":"prior studies conflate these benefits with the implicit correction of mislabeled data (mode shifts), obscuring true effects of soft-labels.","support":"source-stated"},{"text":"In an audit using MNIST and a synthetic variant, the authors report that human soft-labels improved accuracy.","excerpt":"across MNIST and a synthetic variant, re-annotating subsets to extract human uncertainty. By decoupling soft-label supervision from underlying label mode shifts, we show that while human soft-labels do provide accuracy gains","support":"source-stated"},{"text":"The authors say human soft-labels’ larger value was improved calibration on difficult samples and more stable convergence across training runs.","excerpt":"their larger value lies in acting as a regularizer that improves model calibration on difficult samples and promotes stable convergence across training runs.","support":"source-stated"},{"text":"The authors report that models trained on human soft-labels mirrored human uncertainty, unlike models trained on synthetic labels.","excerpt":"models trained on human soft-labels mirror human uncertainty, whereas those trained on synthetic labels fail to align with humans.","support":"source-stated"},{"text":"The authors describe the work as a diagnostic testbed for human-AI uncertainty alignment.","excerpt":"Broadly, this work provides a diagnostic testbed for human-AI uncertainty alignment.","support":"source-stated"}],"limitations":["The supplied evidence is the paper’s abstract and publisher page; it does not provide detailed methods, measurements, or results beyond the summary presented there.","The reported audit covers MNIST and a synthetic variant, so the supplied evidence does not establish whether the findings extend to other datasets or settings."],"mediaId":"abe53f3dfc42ad54e79b90e7b5aa236728787d605b1aded360e9eabbc93d1df3","correctionReason":null,"id":"0179e71f4d5d7f5362533c8adc9e0aba7d3f62136af881944c16233ab1232488","publicationId":"8289903d-91a0-4bf1-8119-fbb425bb4407","version":1,"sourceUrl":"https://arxiv.org/abs/2605.18648","sourceTitle":"An Assessment of Human vs. Model Uncertainty in Soft-Label Learning and Calibration","publisher":"arXiv cs.AI","publishedAt":null,"originalPublished":{"value":"2026-05-18","precision":"day","basis":"publisher_citation_date"},"firstObservedAt":"2026-11-06T05:10:00.855Z","createdAt":"2026-11-08T20:33:19.882Z","historical":true,"entityIds":["arxiv"],"topicIds":["intelligence-and-discovery"],"media":[{"id":"abe53f3dfc42ad54e79b90e7b5aa236728787d605b1aded360e9eabbc93d1df3","kind":"article","url":"https://arxiv.org/abs/2605.18648","sourceUrl":"https://arxiv.org/abs/2605.18648","label":"Read original","provenance":"source_link","verification":{"status":"verified","checkedAt":"2026-11-08T20:32:29.341Z","method":"acquired_source","contentType":"text/html"},"recommendationReason":"Recommended by Jev for understanding this source; checked by Luna."}],"sourceMedia":{"status":"not_found","sourceUrl":"https://arxiv.org/abs/2605.18648","checkedAt":"2026-11-08T21:04:08.245Z","items":[],"note":"No suitable media was found by the supported page extraction."},"editorialDecision":{"outcome":"accept","reasonCodes":["material_development"],"rubricVersion":"stream-admission-v5","mediaId":"abe53f3dfc42ad54e79b90e7b5aa236728787d605b1aded360e9eabbc93d1df3","confidence":0.68,"provider":"typesafe","model":"jev-1.13.0","inputHash":"8ddf3951af3400fa141c8846974b7daeb2142e573021fbf24da089566c861d52","decidedAt":"2026-11-08T20:32:30.090Z"},"provenance":{"textHash":"2841df4b028bebad1af317e4b066ffc2ef5b496343152cb83d7f79559e85e9a2","observationId":"8ffe823e229e7025b68b2b90a53a1c2e6538dcfcbdc3ad5e3814ebd093213694","retrievedAt":"2026-11-08T20:32:29.210Z","kind":"publisher-page","preparer":{"model":"gpt-6-luna","runId":"23d28841-194b-49d3-8af7-b6a14e1f029d"},"reviewer":{"model":"gpt-6-luna","runId":"03bb0137-a079-4eea-8dca-bbd4b387821d"}},"audit":{"status":"passed","reviewedAt":"2026-11-08T20:33:19.882Z","model":"anthropic/claude-opus-5.5"},"sourceAttribution":{"authors":[{"name":"Pavlovic, Maja","url":null,"kind":"unknown"},{"name":"Paun, Silviu","url":null,"kind":"unknown"},{"name":"Poesio, Massimo","url":null,"kind":"unknown"}],"authorRole":"article-author","hosts":[],"status":"stated","authorRoster":{"truncated":false,"omitted":0,"assertionGroupsOmitted":0},"participation":[{"actor":{"name":"Pavlovic, Maja","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Pavlovic, Maja","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2605.18648","locator":"head/meta[name=author|citation_author]"},"assertionId":"bc50690072661891bdd8c327d7688896047d6b0b723df351d8c706ef16d7e835:participation:0","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-08T21:04:08.245Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"bc50690072661891bdd8c327d7688896047d6b0b723df351d8c706ef16d7e835","sourceUrl":"https://arxiv.org/abs/2605.18648","contentSha256":"95f0522532e692633f0b6e169a79e4d06ae427dcc6299640f468489841a66122","textHash":"2841df4b028bebad1af317e4b066ffc2ef5b496343152cb83d7f79559e85e9a2","observedAt":"2026-11-08T21:04:08.245Z"}},{"actor":{"name":"Paun, Silviu","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Paun, Silviu","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2605.18648","locator":"head/meta[name=author|citation_author]"},"assertionId":"bc50690072661891bdd8c327d7688896047d6b0b723df351d8c706ef16d7e835:participation:1","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-08T21:04:08.245Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"bc50690072661891bdd8c327d7688896047d6b0b723df351d8c706ef16d7e835","sourceUrl":"https://arxiv.org/abs/2605.18648","contentSha256":"95f0522532e692633f0b6e169a79e4d06ae427dcc6299640f468489841a66122","textHash":"2841df4b028bebad1af317e4b066ffc2ef5b496343152cb83d7f79559e85e9a2","observedAt":"2026-11-08T21:04:08.245Z"}},{"actor":{"name":"Poesio, Massimo","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Poesio, Massimo","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2605.18648","locator":"head/meta[name=author|citation_author]"},"assertionId":"bc50690072661891bdd8c327d7688896047d6b0b723df351d8c706ef16d7e835:participation:2","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-08T21:04:08.245Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"bc50690072661891bdd8c327d7688896047d6b0b723df351d8c706ef16d7e835","sourceUrl":"https://arxiv.org/abs/2605.18648","contentSha256":"95f0522532e692633f0b6e169a79e4d06ae427dcc6299640f468489841a66122","textHash":"2841df4b028bebad1af317e4b066ffc2ef5b496343152cb83d7f79559e85e9a2","observedAt":"2026-11-08T21:04:08.245Z"}}],"participationCoverage":{"total":3,"returned":3,"complete":true,"fullAssertions":"source-metadata-endpoint"},"distribution":null,"publisher":null,"sourceUrl":"https://arxiv.org/abs/2605.18648","snapshotId":"bc50690072661891bdd8c327d7688896047d6b0b723df351d8c706ef16d7e835","contentSha256":"95f0522532e692633f0b6e169a79e4d06ae427dcc6299640f468489841a66122","textHash":"2841df4b028bebad1af317e4b066ffc2ef5b496343152cb83d7f79559e85e9a2","observedAt":"2026-11-08T21:04:08.245Z","basis":"publisher-page-metadata","sourceLinks":47,"linksTruncated":false,"metadataUrl":"/api/pipeline/stream/0179e71f4d5d7f5362533c8adc9e0aba7d3f62136af881944c16233ab1232488/source"}},"canonicalUrl":"https://escapemeridian.com/readings/pavlovic-paun-and-poesio-argue-human-soft-labels-matter-more-0179e71f","formats":{"html":"/readings/pavlovic-paun-and-poesio-argue-human-soft-labels-matter-more-0179e71f","markdown":"/readings/pavlovic-paun-and-poesio-argue-human-soft-labels-matter-more-0179e71f.md","json":"/readings/pavlovic-paun-and-poesio-argue-human-soft-labels-matter-more-0179e71f.json"},"attribution":"Headline and summary by Escape Meridian, prepared by Luna from the attributed source. A source check is not independent confirmation of the source’s claims.","maintainedGuideLinks":[],"readerContext":{"version":"reader-context-v1","attribution":"Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.","terms":[],"entities":[{"id":"arxiv","name":"arXiv","kind":"organization","context":"arXiv is a research manuscript repository where authors share papers that may not yet have been peer reviewed.","referenceUrls":["https://arxiv.org/","https://info.arxiv.org/help/api/index.html"],"path":"/entities/arxiv","json":"/entities/arxiv.json","markdown":"/entities/arxiv.md"}]},"researchQuestions":[],"withdrawn":false,"references":{"schemaVersion":1,"target":{"type":"reading","id":"0179e71f4d5d7f5362533c8adc9e0aba7d3f62136af881944c16233ab1232488","revision":"8289903d-91a0-4bf1-8119-fbb425bb4407","revisionKind":"publication-id","title":"Pavlovic, Paun and Poesio argue human soft-labels matter more for calibration than accuracy","url":"https://escapemeridian.com/readings/pavlovic-paun-and-poesio-argue-human-soft-labels-matter-more-0179e71f"},"source":{"type":"source","id":"https://arxiv.org/abs/2605.18648","revision":"2841df4b028bebad1af317e4b066ffc2ef5b496343152cb83d7f79559e85e9a2","revisionKind":"sha256-source-text","title":"An Assessment of Human vs. Model Uncertainty in Soft-Label Learning and Calibration","url":"https://arxiv.org/abs/2605.18648"},"status":"available","items":[],"next":null,"scope":"Explicit citations in public readings and public file attachments to this reading. Shared topics are not citations. Other artifact types need an explicit reference adapter."},"revisions":[{"id":"0179e71f4d5d7f5362533c8adc9e0aba7d3f62136af881944c16233ab1232488","title":"Pavlovic, Paun and Poesio argue human soft-labels matter more for calibration than accuracy","argument":"Maja Pavlovic, Silviu Paun and Massimo Poesio argue that human soft-labels—labels that retain uncertainty rather than giving only one answer—may matter less for improving accuracy than for improving calibration: whether a model’s confidence reflects its uncertainty. Their paper reports a controlled audit on MNIST and a synthetic variant, designed to separate the effect of soft-labels from mode shifts, or changes in which label is treated as the main answer. The authors say human soft-labels improved accuracy, but had a larger value as a regularizer, helping calibration on difficult samples and stable convergence across training runs. They also report that dataset cartography showed models trained on human soft-labels mirrored human uncertainty, while models trained on synthetic labels did not align with humans. The distinction matters because earlier studies, they argue, conflated benefits of soft-labels with correction of mislabeled data. The authors present their work as a diagnostic testbed for human-AI uncertainty alignment, not evidence here of performance across other datasets or settings.","topicIds":["intelligence-and-discovery"],"domain":"science","claimType":"attributed reading","limits":"The supplied evidence is the paper’s abstract and publisher page; it does not provide detailed methods, measurements, or results beyond the summary presented there. The reported audit covers MNIST and a synthetic variant, so the supplied evidence does not establish whether the findings extend to other datasets or settings.","speaker":{"name":"Pavlovic, Maja, Paun, Silviu, Poesio, Massimo","roleAtSource":"Article author"},"source":{"title":"An Assessment of Human vs. Model Uncertainty in Soft-Label Learning and Calibration","publisher":"arXiv cs.AI","format":"Article","url":"https://arxiv.org/abs/2605.18648","publishedAt":"2026-05-18","dateBasis":"Date basis not provided."},"evidence":[{"text":"The authors argue earlier work conflated benefits of soft-labels with correction of mislabeled data, obscuring the effects of soft-label supervision itself.","excerpt":"prior studies conflate these benefits with the implicit correction of mislabeled data (mode shifts), obscuring true effects of soft-labels.","url":"https://arxiv.org/abs/2605.18648","relationship":"Source-stated"},{"text":"In an audit using MNIST and a synthetic variant, the authors report that human soft-labels improved accuracy.","excerpt":"across MNIST and a synthetic variant, re-annotating subsets to extract human uncertainty. By decoupling soft-label supervision from underlying label mode shifts, we show that while human soft-labels do provide accuracy gains","url":"https://arxiv.org/abs/2605.18648","relationship":"Source-stated"},{"text":"The authors say human soft-labels’ larger value was improved calibration on difficult samples and more stable convergence across training runs.","excerpt":"their larger value lies in acting as a regularizer that improves model calibration on difficult samples and promotes stable convergence across training runs.","url":"https://arxiv.org/abs/2605.18648","relationship":"Source-stated"},{"text":"The authors report that models trained on human soft-labels mirrored human uncertainty, unlike models trained on synthetic labels.","excerpt":"models trained on human soft-labels mirror human uncertainty, whereas those trained on synthetic labels fail to align with humans.","url":"https://arxiv.org/abs/2605.18648","relationship":"Source-stated"},{"text":"The authors describe the work as a diagnostic testbed for human-AI uncertainty alignment.","excerpt":"Broadly, this work provides a diagnostic testbed for human-AI uncertainty alignment.","url":"https://arxiv.org/abs/2605.18648","relationship":"Source-stated"}],"sourceMedia":{"status":"not_found","sourceUrl":"https://arxiv.org/abs/2605.18648","checkedAt":"2026-11-08T21:04:08.245Z","items":[],"note":"No suitable media was found by the supported page extraction."},"assumptions":[],"openQuestion":{"attribution":"Open question","text":"Do human soft-labels produce the reported calibration and uncertainty-alignment patterns beyond MNIST and the synthetic variant?"},"whatWouldChangeView":{"attribution":"Change condition not recorded","text":""},"editorialComparisons":[],"live":{"title":"Pavlovic, Paun and Poesio argue human soft-labels matter more for calibration than accuracy","summary":"Maja Pavlovic, Silviu Paun and Massimo Poesio argue that human soft-labels—labels that retain uncertainty rather than giving only one answer—may matter less for improving accuracy than for improving calibration: whether a model’s confidence reflects its uncertainty. Their paper reports a controlled audit on MNIST and a synthetic variant, designed to separate the effect of soft-labels from mode shifts, or changes in which label is treated as the main answer. The authors say human soft-labels improved accuracy, but had a larger value as a regularizer, helping calibration on difficult samples and stable convergence across training runs. They also report that dataset cartography showed models trained on human soft-labels mirrored human uncertainty, while models trained on synthetic labels did not align with humans. The distinction matters because earlier studies, they argue, conflated benefits of soft-labels with correction of mislabeled data. The authors present their work as a diagnostic testbed for human-AI uncertainty alignment, not evidence here of performance across other datasets or settings.","question":"Do human soft-labels produce the reported calibration and uncertainty-alignment patterns beyond MNIST and the synthetic variant?","claims":[{"text":"The authors argue earlier work conflated benefits of soft-labels with correction of mislabeled data, obscuring the effects of soft-label supervision itself.","excerpt":"prior studies conflate these benefits with the implicit correction of mislabeled data (mode shifts), obscuring true effects of soft-labels.","support":"source-stated"},{"text":"In an audit using MNIST and a synthetic variant, the authors report that human soft-labels improved accuracy.","excerpt":"across MNIST and a synthetic variant, re-annotating subsets to extract human uncertainty. By decoupling soft-label supervision from underlying label mode shifts, we show that while human soft-labels do provide accuracy gains","support":"source-stated"},{"text":"The authors say human soft-labels’ larger value was improved calibration on difficult samples and more stable convergence across training runs.","excerpt":"their larger value lies in acting as a regularizer that improves model calibration on difficult samples and promotes stable convergence across training runs.","support":"source-stated"},{"text":"The authors report that models trained on human soft-labels mirrored human uncertainty, unlike models trained on synthetic labels.","excerpt":"models trained on human soft-labels mirror human uncertainty, whereas those trained on synthetic labels fail to align with humans.","support":"source-stated"},{"text":"The authors describe the work as a diagnostic testbed for human-AI uncertainty alignment.","excerpt":"Broadly, this work provides a diagnostic testbed for human-AI uncertainty alignment.","support":"source-stated"}],"limitations":["The supplied evidence is the paper’s abstract and publisher page; it does not provide detailed methods, measurements, or results beyond the summary presented there.","The reported audit covers MNIST and a synthetic variant, so the supplied evidence does not establish whether the findings extend to other datasets or settings."],"mediaId":"abe53f3dfc42ad54e79b90e7b5aa236728787d605b1aded360e9eabbc93d1df3","correctionReason":null,"id":"0179e71f4d5d7f5362533c8adc9e0aba7d3f62136af881944c16233ab1232488","publicationId":"8289903d-91a0-4bf1-8119-fbb425bb4407","version":1,"sourceUrl":"https://arxiv.org/abs/2605.18648","sourceTitle":"An Assessment of Human vs. Model Uncertainty in Soft-Label Learning and Calibration","publisher":"arXiv cs.AI","publishedAt":null,"originalPublished":{"value":"2026-05-18","precision":"day","basis":"publisher_citation_date"},"firstObservedAt":"2026-11-06T05:10:00.855Z","createdAt":"2026-11-08T20:33:19.882Z","historical":true,"entityIds":["arxiv"],"topicIds":["intelligence-and-discovery"],"media":[{"id":"abe53f3dfc42ad54e79b90e7b5aa236728787d605b1aded360e9eabbc93d1df3","kind":"article","url":"https://arxiv.org/abs/2605.18648","sourceUrl":"https://arxiv.org/abs/2605.18648","label":"Read original","provenance":"source_link","verification":{"status":"verified","checkedAt":"2026-11-08T20:32:29.341Z","method":"acquired_source","contentType":"text/html"},"recommendationReason":"Recommended by Jev for understanding this source; checked by Luna."}],"sourceMedia":{"status":"not_found","sourceUrl":"https://arxiv.org/abs/2605.18648","checkedAt":"2026-11-08T21:04:08.245Z","items":[],"note":"No suitable media was found by the supported page extraction."},"editorialDecision":{"outcome":"accept","reasonCodes":["material_development"],"rubricVersion":"stream-admission-v5","mediaId":"abe53f3dfc42ad54e79b90e7b5aa236728787d605b1aded360e9eabbc93d1df3","confidence":0.68,"provider":"typesafe","model":"jev-1.13.0","inputHash":"8ddf3951af3400fa141c8846974b7daeb2142e573021fbf24da089566c861d52","decidedAt":"2026-11-08T20:32:30.090Z"},"provenance":{"textHash":"2841df4b028bebad1af317e4b066ffc2ef5b496343152cb83d7f79559e85e9a2","observationId":"8ffe823e229e7025b68b2b90a53a1c2e6538dcfcbdc3ad5e3814ebd093213694","retrievedAt":"2026-11-08T20:32:29.210Z","kind":"publisher-page","preparer":{"model":"gpt-6-luna","runId":"23d28841-194b-49d3-8af7-b6a14e1f029d"},"reviewer":{"model":"gpt-6-luna","runId":"03bb0137-a079-4eea-8dca-bbd4b387821d"}},"audit":{"status":"passed","reviewedAt":"2026-11-08T20:33:19.882Z","model":"anthropic/claude-opus-5.5"},"sourceAttribution":{"authors":[{"name":"Pavlovic, Maja","url":null,"kind":"unknown"},{"name":"Paun, Silviu","url":null,"kind":"unknown"},{"name":"Poesio, Massimo","url":null,"kind":"unknown"}],"authorRole":"article-author","hosts":[],"status":"stated","authorRoster":{"truncated":false,"omitted":0,"assertionGroupsOmitted":0},"participation":[{"actor":{"name":"Pavlovic, Maja","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Pavlovic, Maja","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2605.18648","locator":"head/meta[name=author|citation_author]"},"assertionId":"bc50690072661891bdd8c327d7688896047d6b0b723df351d8c706ef16d7e835:participation:0","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-08T21:04:08.245Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"bc50690072661891bdd8c327d7688896047d6b0b723df351d8c706ef16d7e835","sourceUrl":"https://arxiv.org/abs/2605.18648","contentSha256":"95f0522532e692633f0b6e169a79e4d06ae427dcc6299640f468489841a66122","textHash":"2841df4b028bebad1af317e4b066ffc2ef5b496343152cb83d7f79559e85e9a2","observedAt":"2026-11-08T21:04:08.245Z"}},{"actor":{"name":"Paun, Silviu","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Paun, Silviu","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2605.18648","locator":"head/meta[name=author|citation_author]"},"assertionId":"bc50690072661891bdd8c327d7688896047d6b0b723df351d8c706ef16d7e835:participation:1","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-08T21:04:08.245Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"bc50690072661891bdd8c327d7688896047d6b0b723df351d8c706ef16d7e835","sourceUrl":"https://arxiv.org/abs/2605.18648","contentSha256":"95f0522532e692633f0b6e169a79e4d06ae427dcc6299640f468489841a66122","textHash":"2841df4b028bebad1af317e4b066ffc2ef5b496343152cb83d7f79559e85e9a2","observedAt":"2026-11-08T21:04:08.245Z"}},{"actor":{"name":"Poesio, Massimo","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Poesio, Massimo","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2605.18648","locator":"head/meta[name=author|citation_author]"},"assertionId":"bc50690072661891bdd8c327d7688896047d6b0b723df351d8c706ef16d7e835:participation:2","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-08T21:04:08.245Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"bc50690072661891bdd8c327d7688896047d6b0b723df351d8c706ef16d7e835","sourceUrl":"https://arxiv.org/abs/2605.18648","contentSha256":"95f0522532e692633f0b6e169a79e4d06ae427dcc6299640f468489841a66122","textHash":"2841df4b028bebad1af317e4b066ffc2ef5b496343152cb83d7f79559e85e9a2","observedAt":"2026-11-08T21:04:08.245Z"}}],"participationCoverage":{"total":3,"returned":3,"complete":true,"fullAssertions":"source-metadata-endpoint"},"distribution":null,"publisher":null,"sourceUrl":"https://arxiv.org/abs/2605.18648","snapshotId":"bc50690072661891bdd8c327d7688896047d6b0b723df351d8c706ef16d7e835","contentSha256":"95f0522532e692633f0b6e169a79e4d06ae427dcc6299640f468489841a66122","textHash":"2841df4b028bebad1af317e4b066ffc2ef5b496343152cb83d7f79559e85e9a2","observedAt":"2026-11-08T21:04:08.245Z","basis":"publisher-page-metadata","sourceLinks":47,"linksTruncated":false,"metadataUrl":"/api/pipeline/stream/0179e71f4d5d7f5362533c8adc9e0aba7d3f62136af881944c16233ab1232488/source"}}}],"audits":[{"publicationId":"8289903d-91a0-4bf1-8119-fbb425bb4407","verdict":"passed","findings":[{"dimension":"attribution","status":"pass","explanation":"The draft names the listed authors, Maja Pavlovic, Silviu Paun and Massimo Poesio, and attributes each finding to them as reported or argued claims, not as established facts.","excerpt":"Authors:Maja Pavlovic, Silviu Paun, Massimo Poesio"},{"dimension":"summary","status":"pass","explanation":"The summary keeps the abstract's central argument: soft-labels give some accuracy gain, but their larger value is regularization that improves calibration and stable convergence. It explains soft-labels, calibration and mode shifts in plain language. Regularizer, dataset cartography and MNIST are not explained inline but suit term-library tooltips. The title and the question about generalization beyond MNIST follow from the evidence.","excerpt":"their larger value lies in acting as a regularizer that improves model calibration on difficult samples and promotes stable convergence across training runs."},{"dimension":"date","status":"pass","explanation":"originalPublished is 2026-05-18, matching the v1 submission date. It correctly ignores the later v3 revision of 5 Nov 2026 and the November feed date.","excerpt":"[Submitted on 18 May 2026 (v1), last revised 5 Nov 2026 (this version, v3)]"},{"dimension":"support","status":"pass","explanation":"Every claim's excerpt is an exact quotation from the abstract, and the claims are source-stated. The limitations correctly note that the evidence is only the abstract and covers MNIST and a synthetic variant. The draft does not inflate the findings into real-world reliability claims.","excerpt":"Dataset cartography reveals models trained on human soft-labels mirror human uncertainty, whereas those trained on synthetic labels fail to align with humans."},{"dimension":"media","status":"pass","explanation":"The selected media is the verified direct arXiv abstract page for this paper. It is the original source and the most relevant link.","excerpt":null}],"createdAt":"2026-11-08T20:33:19.882Z","model":"anthropic/claude-opus-5.5"}],"correctionUrl":"https://escapemeridian.com/corrections/"}