{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:45F5CUW3IMYQOQIDNSAHHNCXXO","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"d9dbf981bef549d6867176946de85939839ae7d66dff89e44ec69cc0326de424","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2025-10-03T07:22:36Z","title_canon_sha256":"4bc1d06d560a9fc43b13cd540ea8c56d0986ecce85e5c3fe2893cbb0477bcf0e"},"schema_version":"1.0","source":{"id":"2510.02779","kind":"arxiv","version":4}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2510.02779","created_at":"2026-06-03T01:05:45Z"},{"alias_kind":"arxiv_version","alias_value":"2510.02779v4","created_at":"2026-06-03T01:05:45Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2510.02779","created_at":"2026-06-03T01:05:45Z"},{"alias_kind":"pith_short_12","alias_value":"45F5CUW3IMYQ","created_at":"2026-06-03T01:05:45Z"},{"alias_kind":"pith_short_16","alias_value":"45F5CUW3IMYQOQID","created_at":"2026-06-03T01:05:45Z"},{"alias_kind":"pith_short_8","alias_value":"45F5CUW3","created_at":"2026-06-03T01:05:45Z"}],"graph_snapshots":[{"event_id":"sha256:dcb0970a7f0b2820217cc90bb79f9dac23adf21cca47c7034df55e5eaee7dc21","target":"graph","created_at":"2026-06-03T01:05:45Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":4,"items":[{"attestation":"unclaimed","claim_id":"C1","kind":"strongest_claim","source":"verdict.strongest_claim","status":"machine_extracted","text":"under the assumption that the data are NTK separable from the margin γ, we prove an excess risk rate of Õ(L^6 / (n γ²))"},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"the assumption that the data are NTK separable from the margin γ (stated in the abstract as the condition under which the rate holds)"},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"Gradient descent on deep ReLU networks achieves excess risk rate of order L^6/(n gamma^2) under NTK margin separability via activation pattern control."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"Under NTK separability with margin γ, gradient descent on deep ReLU networks attains an excess risk of Õ(L^6/(n γ²))."}],"snapshot_sha256":"4b9cf7645b6f3a321fff3d62f9b10a2d649b08264e9c9e27f7e46206ad4fab8b"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2510.02779/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Recent advances have significantly improved our understanding of the generalization performance of gradient descent (GD) methods in deep neural networks. A natural and fundamental question is whether GD can achieve generalization rates comparable to the minimax optimal rates established in the kernel setting. Existing results either yield suboptimal rates of $O(1/\\sqrt{n})$, or focus on networks with smooth activation functions, incurring exponential dependence on network depth $L$. In this work, we establish optimal generalization rates for GD with deep ReLU networks by carefully trading off ","authors_text":"Yiming Ying, Yuanfan Li, Yunwen Lei, Zheng-Chu Guo","cross_cats":[],"headline":"Under NTK separability with margin γ, gradient descent on deep ReLU networks attains an excess risk of Õ(L^6/(n γ²)).","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2025-10-03T07:22:36Z","title":"Optimal Rates for Generalization of Gradient Descent for Deep ReLU Classification"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2510.02779","kind":"arxiv","version":4},"verdict":{"created_at":"2026-05-18T10:21:36.489766Z","id":"b659c8cc-a3bc-4133-a4ca-b70e63599fbb","model_set":{"reader":"grok-4.3"},"one_line_summary":"Gradient descent on deep ReLU networks achieves excess risk rate of order L^6/(n gamma^2) under NTK margin separability via activation pattern control.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"Under NTK separability with margin γ, gradient descent on deep ReLU networks attains an excess risk of Õ(L^6/(n γ²)).","strongest_claim":"under the assumption that the data are NTK separable from the margin γ, we prove an excess risk rate of Õ(L^6 / (n γ²))","weakest_assumption":"the assumption that the data are NTK separable from the margin γ (stated in the abstract as the condition under which the rate holds)"}},"verdict_id":"b659c8cc-a3bc-4133-a4ca-b70e63599fbb"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:a1f85fb2ce3eb4f615a44528a49d4d4469ec593e2c44bdee89192ba01f6f3c44","target":"record","created_at":"2026-06-03T01:05:45Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"d9dbf981bef549d6867176946de85939839ae7d66dff89e44ec69cc0326de424","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2025-10-03T07:22:36Z","title_canon_sha256":"4bc1d06d560a9fc43b13cd540ea8c56d0986ecce85e5c3fe2893cbb0477bcf0e"},"schema_version":"1.0","source":{"id":"2510.02779","kind":"arxiv","version":4}},"canonical_sha256":"e74bd152db43310741036c8073b457bbae4282571238749770d7854d1ea081f4","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"e74bd152db43310741036c8073b457bbae4282571238749770d7854d1ea081f4","first_computed_at":"2026-06-03T01:05:45.517619Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-06-03T01:05:45.517619Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"Ze/h3oof7o32FV0R8IUewD4+heKvY0ruVaa7V73Hw4kztoMZ4Ayg36tpBRiD/luRtzae6+dXQqT4ogYcVJtcAQ==","signature_status":"signed_v1","signed_at":"2026-06-03T01:05:45.518159Z","signed_message":"canonical_sha256_bytes"},"source_id":"2510.02779","source_kind":"arxiv","source_version":4}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:a1f85fb2ce3eb4f615a44528a49d4d4469ec593e2c44bdee89192ba01f6f3c44","sha256:dcb0970a7f0b2820217cc90bb79f9dac23adf21cca47c7034df55e5eaee7dc21"],"state_sha256":"84f9a2810cc0a87d5e2961af5e50e0fade98073ba9e7f63968291ff36f3f73a0"}