{"id":"06b199d48a2e8c161e850180a22aef8724a8f5eebb820b568b99122a43d63287","title":"arXiv cs.AI: A Spectral Renormalization-Group View of Learning","argument":"Charles H. Martin proposes a spectral Renormalization-Group view of how layers in deep neural networks converge during training. He says layers often develop heavy-tailed power-law spectra, and that the strongest checkpoints often appear near the power-law exponent α≈2. His statistical hypothesis is that α=2 marks a boundary between stable, if not optimal, learning and overfitting via non-self-averaging. His physical explanation is that this value makes spectral energy gain scale evenly across logarithmic spectral scales, a proposed Wilsonian marginal condition. Martin says the proposed SETOL theory applies an ERG trace-log condition, justified by transformations intended to remove arbitrary layer scale along the proposed RG flow. He connects this spectral RG flow to HTSR, SETOL and optimizers such as Muon, and says it motivates a new optimizer extension, WW-PGD. Escape Meridian’s reading is that the abstract offers these ideas as a proposed account, not a general result: it says neural networks may approach shared spectral structure in the settings studied. The source page’s abstract does not provide the methods or results needed here to assess how broadly that pattern holds.","topicIds":["intelligence-and-discovery"],"domain":"science","claimType":"attributed reading","limits":"The supplied source text gives the abstract and bibliographic page, but not the paper’s methods or detailed results; this evidence alone cannot establish how broadly the proposed pattern holds. The abstract describes α=2 as a statistical hypothesis and a proposed physical condition, and says neural networks may approach shared spectral structure in the settings studied. These are not presented as universal findings.","speaker":{"name":"Martin, Charles H.","roleAtSource":"Article author"},"source":{"title":"A Spectral Renormalization-Group View of Learning","publisher":"arXiv cs.AI","format":"Article","url":"https://arxiv.org/abs/2611.00422","publishedAt":"2026-10-30","dateBasis":"Date basis not provided."},"evidence":[{"text":"Martin proposes using Wilsonian Renormalization-Group ideas to describe how deep neural network layers converge.","excerpt":"We study how the layers in deep neural networks (NNs) converge, using Wilsonian Renormalization-Group (RG) ideas to organize the spectral picture.","url":"https://arxiv.org/abs/2611.00422","relationship":"Source-stated"},{"text":"The abstract says layers often develop heavy-tailed power-law spectra, with the strongest checkpoints often near α≈2.","excerpt":"During training, these layers often develop heavy-tailed power-law (PL) spectra, and the strongest checkpoints often appear near the PL exponent $\\alpha\\approx 2$.","url":"https://arxiv.org/abs/2611.00422","relationship":"Source-stated"},{"text":"Martin proposes both a statistical and a physical explanation for why α=2 could mark a well-trained layer.","excerpt":"The answer is both statistical and physical. Our statistical hypothesis is that $\\alpha=2$ marks the boundary between stable if not optimal learning and overfitting via non-self-averaging.","url":"https://arxiv.org/abs/2611.00422","relationship":"Source-stated"},{"text":"The abstract connects the proposed spectral RG flow to existing theories and optimizers, and says it motivates WW-PGD.","excerpt":"The resulting spectral RG flow connects the HTSR and SETOL theories to optimizers such as Muon, and motivates a new optimizer extension, WW-PGD.","url":"https://arxiv.org/abs/2611.00422","relationship":"Source-stated"},{"text":"The proposed account is limited in the abstract to settings studied; it suggests, rather than establishes, shared spectral structure in neural networks.","excerpt":"Together, these results point toward a spectral Renormalization-Group view of learning, in which NNs may approach shared spectral structure in the settings studied here.","url":"https://arxiv.org/abs/2611.00422","relationship":"Escape Meridian interpretation"}],"sourceMedia":{"status":"not_found","sourceUrl":"https://arxiv.org/abs/2611.00422","checkedAt":"2026-11-10T05:58:06.522Z","items":[],"note":"No suitable media was found by the supported page extraction."},"assumptions":[],"openQuestion":{"attribution":"Open question","text":"What evidence from the settings studied supports the proposed link between α≈2 and well-trained layers, and how consistent is that pattern across those settings?"},"whatWouldChangeView":{"attribution":"Change condition not recorded","text":""},"editorialComparisons":[],"live":{"title":"arXiv cs.AI: A Spectral Renormalization-Group View of Learning","summary":"Charles H. Martin proposes a spectral Renormalization-Group view of how layers in deep neural networks converge during training. He says layers often develop heavy-tailed power-law spectra, and that the strongest checkpoints often appear near the power-law exponent α≈2. His statistical hypothesis is that α=2 marks a boundary between stable, if not optimal, learning and overfitting via non-self-averaging. His physical explanation is that this value makes spectral energy gain scale evenly across logarithmic spectral scales, a proposed Wilsonian marginal condition. Martin says the proposed SETOL theory applies an ERG trace-log condition, justified by transformations intended to remove arbitrary layer scale along the proposed RG flow. He connects this spectral RG flow to HTSR, SETOL and optimizers such as Muon, and says it motivates a new optimizer extension, WW-PGD. Escape Meridian’s reading is that the abstract offers these ideas as a proposed account, not a general result: it says neural networks may approach shared spectral structure in the settings studied. The source page’s abstract does not provide the methods or results needed here to assess how broadly that pattern holds.","question":"What evidence from the settings studied supports the proposed link between α≈2 and well-trained layers, and how consistent is that pattern across those settings?","claims":[{"text":"Martin proposes using Wilsonian Renormalization-Group ideas to describe how deep neural network layers converge.","excerpt":"We study how the layers in deep neural networks (NNs) converge, using Wilsonian Renormalization-Group (RG) ideas to organize the spectral picture.","support":"source-stated"},{"text":"The abstract says layers often develop heavy-tailed power-law spectra, with the strongest checkpoints often near α≈2.","excerpt":"During training, these layers often develop heavy-tailed power-law (PL) spectra, and the strongest checkpoints often appear near the PL exponent $\\alpha\\approx 2$.","support":"source-stated"},{"text":"Martin proposes both a statistical and a physical explanation for why α=2 could mark a well-trained layer.","excerpt":"The answer is both statistical and physical. Our statistical hypothesis is that $\\alpha=2$ marks the boundary between stable if not optimal learning and overfitting via non-self-averaging.","support":"source-stated"},{"text":"The abstract connects the proposed spectral RG flow to existing theories and optimizers, and says it motivates WW-PGD.","excerpt":"The resulting spectral RG flow connects the HTSR and SETOL theories to optimizers such as Muon, and motivates a new optimizer extension, WW-PGD.","support":"source-stated"},{"text":"The proposed account is limited in the abstract to settings studied; it suggests, rather than establishes, shared spectral structure in neural networks.","excerpt":"Together, these results point toward a spectral Renormalization-Group view of learning, in which NNs may approach shared spectral structure in the settings studied here.","support":"interpretation"}],"limitations":["The supplied source text gives the abstract and bibliographic page, but not the paper’s methods or detailed results; this evidence alone cannot establish how broadly the proposed pattern holds.","The abstract describes α=2 as a statistical hypothesis and a proposed physical condition, and says neural networks may approach shared spectral structure in the settings studied. These are not presented as universal findings."],"mediaId":"88b138615c14fff05e66ddb68652f5ff95a45f47c39a7ba8c616018383476f0c","correctionReason":null,"id":"06b199d48a2e8c161e850180a22aef8724a8f5eebb820b568b99122a43d63287","publicationId":"336135a1-72b2-4681-85ee-1c7521e68d63","version":1,"sourceUrl":"https://arxiv.org/abs/2611.00422","sourceTitle":"A Spectral Renormalization-Group View of Learning","publisher":"arXiv cs.AI","publishedAt":null,"originalPublished":{"value":"2026-10-30","precision":"day","basis":"publisher_citation_date"},"firstObservedAt":"2026-11-03T08:51:34.123Z","createdAt":"2026-11-10T04:54:10.244Z","historical":true,"entityIds":["arxiv"],"topicIds":["intelligence-and-discovery"],"media":[{"id":"88b138615c14fff05e66ddb68652f5ff95a45f47c39a7ba8c616018383476f0c","kind":"article","url":"https://arxiv.org/abs/2611.00422","sourceUrl":"https://arxiv.org/abs/2611.00422","label":"Read original","provenance":"source_link","verification":{"status":"verified","checkedAt":"2026-11-04T05:01:23.893Z","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/2611.00422","checkedAt":"2026-11-10T05:58:06.522Z","items":[],"note":"No suitable media was found by the supported page extraction."},"editorialDecision":{"outcome":"accept","reasonCodes":["useful_perspective"],"rubricVersion":"stream-admission-v5","mediaId":"88b138615c14fff05e66ddb68652f5ff95a45f47c39a7ba8c616018383476f0c","confidence":0.54,"provider":"typesafe","model":"jev-1.13.0","inputHash":"d66947c5bb9e017c313dd2dbb6f1f49c8b77f1cd87ff7d753ee2f8ebb8378915","decidedAt":"2026-11-04T05:01:25.570Z","redecided":{"from":"stream-admission-v5","under":"stream-admission-v5+confidence-0.5"}},"provenance":{"textHash":"904551d5458a421a3cc977960eb31f6acd0ae1ee2a123d0a9677be7a13181c9d","observationId":"272bb48c5442bbeade6ffdd6969d8911883887fd08d90b96e7995be4fb89a210","retrievedAt":"2026-11-04T05:01:23.809Z","kind":"publisher-page","preparer":{"model":"gpt-6-luna","runId":"29f001b0-15ab-4eea-840f-bcbc202058ad"},"reviewer":{"model":"gpt-6-luna","runId":"75903f3c-5ee7-4c2a-8be3-05849f567485"}},"audit":{"status":"uncertain","reviewedAt":"2026-11-10T04:54:10.244Z","model":"anthropic/claude-opus-5.5"},"sourceAttribution":{"authors":[{"name":"Martin, Charles H.","url":null,"kind":"unknown"}],"authorRole":"article-author","hosts":[],"status":"stated","authorRoster":{"truncated":false,"omitted":0,"assertionGroupsOmitted":0},"participation":[{"actor":{"name":"Martin, Charles H.","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Martin, Charles H.","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.00422","locator":"head/meta[name=author|citation_author]"},"assertionId":"c1bb5e9af32e747bf95709aac68f4368a1e43513c30fdc62363a76d2f37e3025:participation:0","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-04T05:01:23.809Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"c1bb5e9af32e747bf95709aac68f4368a1e43513c30fdc62363a76d2f37e3025","sourceUrl":"https://arxiv.org/abs/2611.00422","contentSha256":"53a0dd95661361cba5d44efdd997e589671fa2bb3c933f17c157286c11ff612e","textHash":"904551d5458a421a3cc977960eb31f6acd0ae1ee2a123d0a9677be7a13181c9d","observedAt":"2026-11-04T05:01:23.809Z"}}],"participationCoverage":{"total":1,"returned":1,"complete":true,"fullAssertions":"source-metadata-endpoint"},"distribution":null,"publisher":null,"sourceUrl":"https://arxiv.org/abs/2611.00422","snapshotId":"c1bb5e9af32e747bf95709aac68f4368a1e43513c30fdc62363a76d2f37e3025","contentSha256":"53a0dd95661361cba5d44efdd997e589671fa2bb3c933f17c157286c11ff612e","textHash":"904551d5458a421a3cc977960eb31f6acd0ae1ee2a123d0a9677be7a13181c9d","observedAt":"2026-11-04T05:01:23.809Z","basis":"publisher-page-metadata","sourceLinks":43,"linksTruncated":false,"metadataUrl":"/api/pipeline/stream/06b199d48a2e8c161e850180a22aef8724a8f5eebb820b568b99122a43d63287/source"}},"canonicalUrl":"https://escapemeridian.com/readings/arxiv-cs-ai-a-spectral-renormalization-group-view-of-learning-06b199d4","formats":{"html":"/readings/arxiv-cs-ai-a-spectral-renormalization-group-view-of-learning-06b199d4","markdown":"/readings/arxiv-cs-ai-a-spectral-renormalization-group-view-of-learning-06b199d4.md","json":"/readings/arxiv-cs-ai-a-spectral-renormalization-group-view-of-learning-06b199d4.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":[{"id":"term-deep-learning","name":"Deep learning","explanation":"Machine learning that uses many layers of simple calculations; the basis of most current AI.","referenceUrl":"https://en.wikipedia.org/wiki/Deep_learning"},{"id":"term-overfitting","name":"Overfitting","explanation":"When a model learns its training examples too closely and performs worse on new data.","referenceUrl":"https://developers.google.com/machine-learning/glossary#overfitting"},{"id":"term-neural-network","name":"Neural network","explanation":"A system of connected simple calculations, loosely inspired by the brain, that learns patterns from data.","referenceUrl":"https://developers.google.com/machine-learning/glossary#neural-network"}],"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":"06b199d48a2e8c161e850180a22aef8724a8f5eebb820b568b99122a43d63287","revision":"336135a1-72b2-4681-85ee-1c7521e68d63","revisionKind":"publication-id","title":"arXiv cs.AI: A Spectral Renormalization-Group View of Learning","url":"https://escapemeridian.com/readings/arxiv-cs-ai-a-spectral-renormalization-group-view-of-learning-06b199d4"},"source":{"type":"source","id":"https://arxiv.org/abs/2611.00422","revision":"904551d5458a421a3cc977960eb31f6acd0ae1ee2a123d0a9677be7a13181c9d","revisionKind":"sha256-source-text","title":"A Spectral Renormalization-Group View of Learning","url":"https://arxiv.org/abs/2611.00422"},"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":"06b199d48a2e8c161e850180a22aef8724a8f5eebb820b568b99122a43d63287","title":"arXiv cs.AI: A Spectral Renormalization-Group View of Learning","argument":"Charles H. Martin proposes a spectral Renormalization-Group view of how layers in deep neural networks converge during training. He says layers often develop heavy-tailed power-law spectra, and that the strongest checkpoints often appear near the power-law exponent α≈2. His statistical hypothesis is that α=2 marks a boundary between stable, if not optimal, learning and overfitting via non-self-averaging. His physical explanation is that this value makes spectral energy gain scale evenly across logarithmic spectral scales, a proposed Wilsonian marginal condition. Martin says the proposed SETOL theory applies an ERG trace-log condition, justified by transformations intended to remove arbitrary layer scale along the proposed RG flow. He connects this spectral RG flow to HTSR, SETOL and optimizers such as Muon, and says it motivates a new optimizer extension, WW-PGD. Escape Meridian’s reading is that the abstract offers these ideas as a proposed account, not a general result: it says neural networks may approach shared spectral structure in the settings studied. The source page’s abstract does not provide the methods or results needed here to assess how broadly that pattern holds.","topicIds":["intelligence-and-discovery"],"domain":"science","claimType":"attributed reading","limits":"The supplied source text gives the abstract and bibliographic page, but not the paper’s methods or detailed results; this evidence alone cannot establish how broadly the proposed pattern holds. The abstract describes α=2 as a statistical hypothesis and a proposed physical condition, and says neural networks may approach shared spectral structure in the settings studied. These are not presented as universal findings.","speaker":{"name":"Martin, Charles H.","roleAtSource":"Article author"},"source":{"title":"A Spectral Renormalization-Group View of Learning","publisher":"arXiv cs.AI","format":"Article","url":"https://arxiv.org/abs/2611.00422","publishedAt":"2026-10-30","dateBasis":"Date basis not provided."},"evidence":[{"text":"Martin proposes using Wilsonian Renormalization-Group ideas to describe how deep neural network layers converge.","excerpt":"We study how the layers in deep neural networks (NNs) converge, using Wilsonian Renormalization-Group (RG) ideas to organize the spectral picture.","url":"https://arxiv.org/abs/2611.00422","relationship":"Source-stated"},{"text":"The abstract says layers often develop heavy-tailed power-law spectra, with the strongest checkpoints often near α≈2.","excerpt":"During training, these layers often develop heavy-tailed power-law (PL) spectra, and the strongest checkpoints often appear near the PL exponent $\\alpha\\approx 2$.","url":"https://arxiv.org/abs/2611.00422","relationship":"Source-stated"},{"text":"Martin proposes both a statistical and a physical explanation for why α=2 could mark a well-trained layer.","excerpt":"The answer is both statistical and physical. Our statistical hypothesis is that $\\alpha=2$ marks the boundary between stable if not optimal learning and overfitting via non-self-averaging.","url":"https://arxiv.org/abs/2611.00422","relationship":"Source-stated"},{"text":"The abstract connects the proposed spectral RG flow to existing theories and optimizers, and says it motivates WW-PGD.","excerpt":"The resulting spectral RG flow connects the HTSR and SETOL theories to optimizers such as Muon, and motivates a new optimizer extension, WW-PGD.","url":"https://arxiv.org/abs/2611.00422","relationship":"Source-stated"},{"text":"The proposed account is limited in the abstract to settings studied; it suggests, rather than establishes, shared spectral structure in neural networks.","excerpt":"Together, these results point toward a spectral Renormalization-Group view of learning, in which NNs may approach shared spectral structure in the settings studied here.","url":"https://arxiv.org/abs/2611.00422","relationship":"Escape Meridian interpretation"}],"sourceMedia":{"status":"not_found","sourceUrl":"https://arxiv.org/abs/2611.00422","checkedAt":"2026-11-10T05:58:06.522Z","items":[],"note":"No suitable media was found by the supported page extraction."},"assumptions":[],"openQuestion":{"attribution":"Open question","text":"What evidence from the settings studied supports the proposed link between α≈2 and well-trained layers, and how consistent is that pattern across those settings?"},"whatWouldChangeView":{"attribution":"Change condition not recorded","text":""},"editorialComparisons":[],"live":{"title":"arXiv cs.AI: A Spectral Renormalization-Group View of Learning","summary":"Charles H. Martin proposes a spectral Renormalization-Group view of how layers in deep neural networks converge during training. He says layers often develop heavy-tailed power-law spectra, and that the strongest checkpoints often appear near the power-law exponent α≈2. His statistical hypothesis is that α=2 marks a boundary between stable, if not optimal, learning and overfitting via non-self-averaging. His physical explanation is that this value makes spectral energy gain scale evenly across logarithmic spectral scales, a proposed Wilsonian marginal condition. Martin says the proposed SETOL theory applies an ERG trace-log condition, justified by transformations intended to remove arbitrary layer scale along the proposed RG flow. He connects this spectral RG flow to HTSR, SETOL and optimizers such as Muon, and says it motivates a new optimizer extension, WW-PGD. Escape Meridian’s reading is that the abstract offers these ideas as a proposed account, not a general result: it says neural networks may approach shared spectral structure in the settings studied. The source page’s abstract does not provide the methods or results needed here to assess how broadly that pattern holds.","question":"What evidence from the settings studied supports the proposed link between α≈2 and well-trained layers, and how consistent is that pattern across those settings?","claims":[{"text":"Martin proposes using Wilsonian Renormalization-Group ideas to describe how deep neural network layers converge.","excerpt":"We study how the layers in deep neural networks (NNs) converge, using Wilsonian Renormalization-Group (RG) ideas to organize the spectral picture.","support":"source-stated"},{"text":"The abstract says layers often develop heavy-tailed power-law spectra, with the strongest checkpoints often near α≈2.","excerpt":"During training, these layers often develop heavy-tailed power-law (PL) spectra, and the strongest checkpoints often appear near the PL exponent $\\alpha\\approx 2$.","support":"source-stated"},{"text":"Martin proposes both a statistical and a physical explanation for why α=2 could mark a well-trained layer.","excerpt":"The answer is both statistical and physical. Our statistical hypothesis is that $\\alpha=2$ marks the boundary between stable if not optimal learning and overfitting via non-self-averaging.","support":"source-stated"},{"text":"The abstract connects the proposed spectral RG flow to existing theories and optimizers, and says it motivates WW-PGD.","excerpt":"The resulting spectral RG flow connects the HTSR and SETOL theories to optimizers such as Muon, and motivates a new optimizer extension, WW-PGD.","support":"source-stated"},{"text":"The proposed account is limited in the abstract to settings studied; it suggests, rather than establishes, shared spectral structure in neural networks.","excerpt":"Together, these results point toward a spectral Renormalization-Group view of learning, in which NNs may approach shared spectral structure in the settings studied here.","support":"interpretation"}],"limitations":["The supplied source text gives the abstract and bibliographic page, but not the paper’s methods or detailed results; this evidence alone cannot establish how broadly the proposed pattern holds.","The abstract describes α=2 as a statistical hypothesis and a proposed physical condition, and says neural networks may approach shared spectral structure in the settings studied. These are not presented as universal findings."],"mediaId":"88b138615c14fff05e66ddb68652f5ff95a45f47c39a7ba8c616018383476f0c","correctionReason":null,"id":"06b199d48a2e8c161e850180a22aef8724a8f5eebb820b568b99122a43d63287","publicationId":"336135a1-72b2-4681-85ee-1c7521e68d63","version":1,"sourceUrl":"https://arxiv.org/abs/2611.00422","sourceTitle":"A Spectral Renormalization-Group View of Learning","publisher":"arXiv cs.AI","publishedAt":null,"originalPublished":{"value":"2026-10-30","precision":"day","basis":"publisher_citation_date"},"firstObservedAt":"2026-11-03T08:51:34.123Z","createdAt":"2026-11-10T04:54:10.244Z","historical":true,"entityIds":["arxiv"],"topicIds":["intelligence-and-discovery"],"media":[{"id":"88b138615c14fff05e66ddb68652f5ff95a45f47c39a7ba8c616018383476f0c","kind":"article","url":"https://arxiv.org/abs/2611.00422","sourceUrl":"https://arxiv.org/abs/2611.00422","label":"Read original","provenance":"source_link","verification":{"status":"verified","checkedAt":"2026-11-04T05:01:23.893Z","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/2611.00422","checkedAt":"2026-11-10T05:58:06.522Z","items":[],"note":"No suitable media was found by the supported page extraction."},"editorialDecision":{"outcome":"accept","reasonCodes":["useful_perspective"],"rubricVersion":"stream-admission-v5","mediaId":"88b138615c14fff05e66ddb68652f5ff95a45f47c39a7ba8c616018383476f0c","confidence":0.54,"provider":"typesafe","model":"jev-1.13.0","inputHash":"d66947c5bb9e017c313dd2dbb6f1f49c8b77f1cd87ff7d753ee2f8ebb8378915","decidedAt":"2026-11-04T05:01:25.570Z","redecided":{"from":"stream-admission-v5","under":"stream-admission-v5+confidence-0.5"}},"provenance":{"textHash":"904551d5458a421a3cc977960eb31f6acd0ae1ee2a123d0a9677be7a13181c9d","observationId":"272bb48c5442bbeade6ffdd6969d8911883887fd08d90b96e7995be4fb89a210","retrievedAt":"2026-11-04T05:01:23.809Z","kind":"publisher-page","preparer":{"model":"gpt-6-luna","runId":"29f001b0-15ab-4eea-840f-bcbc202058ad"},"reviewer":{"model":"gpt-6-luna","runId":"75903f3c-5ee7-4c2a-8be3-05849f567485"}},"audit":{"status":"uncertain","reviewedAt":"2026-11-10T04:54:10.244Z","model":"anthropic/claude-opus-5.5"},"sourceAttribution":{"authors":[{"name":"Martin, Charles H.","url":null,"kind":"unknown"}],"authorRole":"article-author","hosts":[],"status":"stated","authorRoster":{"truncated":false,"omitted":0,"assertionGroupsOmitted":0},"participation":[{"actor":{"name":"Martin, Charles H.","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Martin, Charles H.","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.00422","locator":"head/meta[name=author|citation_author]"},"assertionId":"c1bb5e9af32e747bf95709aac68f4368a1e43513c30fdc62363a76d2f37e3025:participation:0","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-04T05:01:23.809Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"c1bb5e9af32e747bf95709aac68f4368a1e43513c30fdc62363a76d2f37e3025","sourceUrl":"https://arxiv.org/abs/2611.00422","contentSha256":"53a0dd95661361cba5d44efdd997e589671fa2bb3c933f17c157286c11ff612e","textHash":"904551d5458a421a3cc977960eb31f6acd0ae1ee2a123d0a9677be7a13181c9d","observedAt":"2026-11-04T05:01:23.809Z"}}],"participationCoverage":{"total":1,"returned":1,"complete":true,"fullAssertions":"source-metadata-endpoint"},"distribution":null,"publisher":null,"sourceUrl":"https://arxiv.org/abs/2611.00422","snapshotId":"c1bb5e9af32e747bf95709aac68f4368a1e43513c30fdc62363a76d2f37e3025","contentSha256":"53a0dd95661361cba5d44efdd997e589671fa2bb3c933f17c157286c11ff612e","textHash":"904551d5458a421a3cc977960eb31f6acd0ae1ee2a123d0a9677be7a13181c9d","observedAt":"2026-11-04T05:01:23.809Z","basis":"publisher-page-metadata","sourceLinks":43,"linksTruncated":false,"metadataUrl":"/api/pipeline/stream/06b199d48a2e8c161e850180a22aef8724a8f5eebb820b568b99122a43d63287/source"}}}],"audits":[{"publicationId":"336135a1-72b2-4681-85ee-1c7521e68d63","verdict":"uncertain","findings":[{"dimension":"attribution","status":"pass","explanation":"The reading correctly attributes the paper and its hypotheses to Charles H. Martin, the sole listed author. It keeps the claims framed as his proposals, and the EM interpretation is clearly labeled.","excerpt":"Authors:Charles H. Martin"},{"dimension":"summary","status":"uncertain","explanation":"The summary faithfully tracks the abstract, but its central terms go unexplained for a general reader: Renormalization-Group, layer spectra, power-law exponent α, checkpoints, non-self-averaging, HTSR, SETOL, ERG, Muon and WW-PGD. A non-specialist cannot tell what a layer's spectrum is, or why α≈2 marking well-trained layers would matter. Smallest repair: add plain-language glosses or tooltips for these terms, plus one sentence on the stated stakes. One example is telling a well-trained layer from overfitting, as the abstract frames it. Do not add an invented significance claim.","excerpt":"This pattern raises a simple question: why should this value mark a well-trained layer?"},{"dimension":"date","status":"pass","explanation":"The originalPublished date of 2026-10-30 matches the v1 submission date on the arXiv page. The 2026-11-04 date is a later feed date and is not the publication date.","excerpt":"[Submitted on 30 Oct 2026]"},{"dimension":"support","status":"pass","explanation":"Each claim is quoted from or closely paraphrased from the abstract. The hedges 'often', 'hypothesis', 'proposed' and 'may ... in the settings studied here' are preserved. The interpretation claim is labeled and stays within the source. One wording shifts slightly: the abstract says the flow connects the HTSR and SETOL theories to optimizers, while the draft lists all three as connection targets. This is minor and not a material distortion. No benefits or costs are invented, and the question follows from the evidence.","excerpt":"The resulting spectral RG flow connects the HTSR and SETOL theories to optimizers such as Muon, and motivates a new optimizer extension, WW-PGD."},{"dimension":"media","status":"pass","explanation":"The selected media is the verified direct arXiv abstract page for this paper. It matches the source URL and is the most direct way to access the original.","excerpt":null}],"createdAt":"2026-11-10T04:54:10.244Z","model":"anthropic/claude-opus-5.5"}],"correctionUrl":"https://escapemeridian.com/corrections/"}