{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2019:6JZ2B2CWOTHRM2G64BWPV4IO73","short_pith_number":"pith:6JZ2B2CW","schema_version":"1.0","canonical_sha256":"f273a0e85674cf1668dee06cfaf10efeec38fe392740556e1fd13648274ae10f","source":{"kind":"arxiv","id":"1910.04284","version":5},"attestation_state":"computed","paper":{"title":"Improved Sample Complexities for Deep Networks and Robust Classification via an All-Layer Margin","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Colin Wei, Tengyu Ma","submitted_at":"2019-10-09T22:45:45Z","abstract_excerpt":"For linear classifiers, the relationship between (normalized) output margin and generalization is captured in a clear and simple bound -- a large output margin implies good generalization. Unfortunately, for deep models, this relationship is less clear: existing analyses of the output margin give complicated bounds which sometimes depend exponentially on depth. In this work, we propose to instead analyze a new notion of margin, which we call the \"all-layer margin.\" Our analysis reveals that the all-layer margin has a clear and direct relationship with generalization for deep models. This enabl"},"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":"1910.04284","kind":"arxiv","version":5},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2019-10-09T22:45:45Z","cross_cats_sorted":["stat.ML"],"title_canon_sha256":"48596b4d9f6de2d80b60bb155c930392310b9bd0f8fa4147ee0cd3880898d10e","abstract_canon_sha256":"c552d705db6d53b87ef90a4d069cad4a4f774203e77d0c5054fd95e133b88af2"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T02:49:49.653835Z","signature_b64":"qXNaGYiHyCTtxPT/Op7uSeRTROO/TbRto9AS1AW6fOU7WMvrizEQ+Mw/Zutzq6nbFhtIk6ATuOLucLctYJLBDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f273a0e85674cf1668dee06cfaf10efeec38fe392740556e1fd13648274ae10f","last_reissued_at":"2026-07-05T02:49:49.653281Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T02:49:49.653281Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Improved Sample Complexities for Deep Networks and Robust Classification via an All-Layer Margin","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Colin Wei, Tengyu Ma","submitted_at":"2019-10-09T22:45:45Z","abstract_excerpt":"For linear classifiers, the relationship between (normalized) output margin and generalization is captured in a clear and simple bound -- a large output margin implies good generalization. Unfortunately, for deep models, this relationship is less clear: existing analyses of the output margin give complicated bounds which sometimes depend exponentially on depth. In this work, we propose to instead analyze a new notion of margin, which we call the \"all-layer margin.\" Our analysis reveals that the all-layer margin has a clear and direct relationship with generalization for deep models. This enabl"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1910.04284","kind":"arxiv","version":5},"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/1910.04284/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":"1910.04284","created_at":"2026-07-05T02:49:49.653345+00:00"},{"alias_kind":"arxiv_version","alias_value":"1910.04284v5","created_at":"2026-07-05T02:49:49.653345+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1910.04284","created_at":"2026-07-05T02:49:49.653345+00:00"},{"alias_kind":"pith_short_12","alias_value":"6JZ2B2CWOTHR","created_at":"2026-07-05T02:49:49.653345+00:00"},{"alias_kind":"pith_short_16","alias_value":"6JZ2B2CWOTHRM2G6","created_at":"2026-07-05T02:49:49.653345+00:00"},{"alias_kind":"pith_short_8","alias_value":"6JZ2B2CW","created_at":"2026-07-05T02:49:49.653345+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.09658","citing_title":"Muon Learns More Robust and Transferable Features than Adam","ref_index":139,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/6JZ2B2CWOTHRM2G64BWPV4IO73","json":"https://pith.science/pith/6JZ2B2CWOTHRM2G64BWPV4IO73.json","graph_json":"https://pith.science/api/pith-number/6JZ2B2CWOTHRM2G64BWPV4IO73/graph.json","events_json":"https://pith.science/api/pith-number/6JZ2B2CWOTHRM2G64BWPV4IO73/events.json","paper":"https://pith.science/paper/6JZ2B2CW"},"agent_actions":{"view_html":"https://pith.science/pith/6JZ2B2CWOTHRM2G64BWPV4IO73","download_json":"https://pith.science/pith/6JZ2B2CWOTHRM2G64BWPV4IO73.json","view_paper":"https://pith.science/paper/6JZ2B2CW","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1910.04284&json=true","fetch_graph":"https://pith.science/api/pith-number/6JZ2B2CWOTHRM2G64BWPV4IO73/graph.json","fetch_events":"https://pith.science/api/pith-number/6JZ2B2CWOTHRM2G64BWPV4IO73/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/6JZ2B2CWOTHRM2G64BWPV4IO73/action/timestamp_anchor","attest_storage":"https://pith.science/pith/6JZ2B2CWOTHRM2G64BWPV4IO73/action/storage_attestation","attest_author":"https://pith.science/pith/6JZ2B2CWOTHRM2G64BWPV4IO73/action/author_attestation","sign_citation":"https://pith.science/pith/6JZ2B2CWOTHRM2G64BWPV4IO73/action/citation_signature","submit_replication":"https://pith.science/pith/6JZ2B2CWOTHRM2G64BWPV4IO73/action/replication_record"}},"created_at":"2026-07-05T02:49:49.653345+00:00","updated_at":"2026-07-05T02:49:49.653345+00:00"}