{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:W33VOY6JGCT4T7FGANDRKL5SG5","short_pith_number":"pith:W33VOY6J","schema_version":"1.0","canonical_sha256":"b6f75763c930a7c9fca60347152fb237755a874955a1c6f5878bb116e54b4397","source":{"kind":"arxiv","id":"2606.20469","version":1},"attestation_state":"computed","paper":{"title":"Fisher-Geometric Sharpness and the Implicit Bias of SGD toward Flat Minima","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CG"],"primary_cat":"cs.LG","authors_text":"Hemen Dutta, Kumaresh Sarmah, Md Sakir Ahmed","submitted_at":"2026-06-18T16:48:18Z","abstract_excerpt":"A widely held intuition in deep learning is that stochastic gradient descent (SGD) implicitly favors flat minima and that flat minima generalize better, but standard Euclidean measures of flatness such as the trace or maximum eigenvalue of the loss Hessian are not invariant under reparametrizations that preserve the network function, which undermines the theoretical foundations of this narrative. In this study we resolve this issue by grounding flatness in the Riemannian geometry of the statistical manifold induced by the Fisher Information Matrix (FIM). We define Riemannian sharpness mathemat"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2606.20469","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-06-18T16:48:18Z","cross_cats_sorted":["cs.CG"],"title_canon_sha256":"1df6b25462d2978963d831a2c2b381f734726012302dc9d9ebf4fa470c653e9f","abstract_canon_sha256":"1add1cd60f12d80f3bd0953fae2e2ccfc68937f30d9864e8bb30cc429a83292c"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-19T16:13:13.134541Z","signature_b64":"8FqlXAwCPRVPxnnIeYh7Wj9/Q3CzNJRu9AxsjkrkJsyXb4mQJG25g37R6Sd2XFsU1UBw46bFVjXXIbyAx5SvDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b6f75763c930a7c9fca60347152fb237755a874955a1c6f5878bb116e54b4397","last_reissued_at":"2026-06-19T16:13:13.134120Z","signature_status":"signed_v1","first_computed_at":"2026-06-19T16:13:13.134120Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Fisher-Geometric Sharpness and the Implicit Bias of SGD toward Flat Minima","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CG"],"primary_cat":"cs.LG","authors_text":"Hemen Dutta, Kumaresh Sarmah, Md Sakir Ahmed","submitted_at":"2026-06-18T16:48:18Z","abstract_excerpt":"A widely held intuition in deep learning is that stochastic gradient descent (SGD) implicitly favors flat minima and that flat minima generalize better, but standard Euclidean measures of flatness such as the trace or maximum eigenvalue of the loss Hessian are not invariant under reparametrizations that preserve the network function, which undermines the theoretical foundations of this narrative. In this study we resolve this issue by grounding flatness in the Riemannian geometry of the statistical manifold induced by the Fisher Information Matrix (FIM). We define Riemannian sharpness mathemat"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.20469","kind":"arxiv","version":1},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2606.20469/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2606.20469","created_at":"2026-06-19T16:13:13.134185+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.20469v1","created_at":"2026-06-19T16:13:13.134185+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.20469","created_at":"2026-06-19T16:13:13.134185+00:00"},{"alias_kind":"pith_short_12","alias_value":"W33VOY6JGCT4","created_at":"2026-06-19T16:13:13.134185+00:00"},{"alias_kind":"pith_short_16","alias_value":"W33VOY6JGCT4T7FG","created_at":"2026-06-19T16:13:13.134185+00:00"},{"alias_kind":"pith_short_8","alias_value":"W33VOY6J","created_at":"2026-06-19T16:13:13.134185+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/W33VOY6JGCT4T7FGANDRKL5SG5","json":"https://pith.science/pith/W33VOY6JGCT4T7FGANDRKL5SG5.json","graph_json":"https://pith.science/api/pith-number/W33VOY6JGCT4T7FGANDRKL5SG5/graph.json","events_json":"https://pith.science/api/pith-number/W33VOY6JGCT4T7FGANDRKL5SG5/events.json","paper":"https://pith.science/paper/W33VOY6J"},"agent_actions":{"view_html":"https://pith.science/pith/W33VOY6JGCT4T7FGANDRKL5SG5","download_json":"https://pith.science/pith/W33VOY6JGCT4T7FGANDRKL5SG5.json","view_paper":"https://pith.science/paper/W33VOY6J","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.20469&json=true","fetch_graph":"https://pith.science/api/pith-number/W33VOY6JGCT4T7FGANDRKL5SG5/graph.json","fetch_events":"https://pith.science/api/pith-number/W33VOY6JGCT4T7FGANDRKL5SG5/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/W33VOY6JGCT4T7FGANDRKL5SG5/action/timestamp_anchor","attest_storage":"https://pith.science/pith/W33VOY6JGCT4T7FGANDRKL5SG5/action/storage_attestation","attest_author":"https://pith.science/pith/W33VOY6JGCT4T7FGANDRKL5SG5/action/author_attestation","sign_citation":"https://pith.science/pith/W33VOY6JGCT4T7FGANDRKL5SG5/action/citation_signature","submit_replication":"https://pith.science/pith/W33VOY6JGCT4T7FGANDRKL5SG5/action/replication_record"}},"created_at":"2026-06-19T16:13:13.134185+00:00","updated_at":"2026-06-19T16:13:13.134185+00:00"}