{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2019:7ENGHTO462KM5HICIS45ZQ6LWY","short_pith_number":"pith:7ENGHTO4","schema_version":"1.0","canonical_sha256":"f91a63cddcf694ce9d0244b9dcc3cbb60977231e32c33045b2307bd82fffdd4d","source":{"kind":"arxiv","id":"1905.07325","version":1},"attestation_state":"computed","paper":{"title":"Lexicographic and Depth-Sensitive Margins in Homogeneous and Non-Homogeneous Deep Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"stat.ML","authors_text":"Daniel Soudry, Jason D. Lee, Mor Shpigel Nacson, Nathan Srebro, Suriya Gunasekar","submitted_at":"2019-05-17T15:24:33Z","abstract_excerpt":"With an eye toward understanding complexity control in deep learning, we study how infinitesimal regularization or gradient descent optimization lead to margin maximizing solutions in both homogeneous and non-homogeneous models, extending previous work that focused on infinitesimal regularization only in homogeneous models. To this end we study the limit of loss minimization with a diverging norm constraint (the \"constrained path\"), relate it to the limit of a \"margin path\" and characterize the resulting solution. For non-homogeneous ensemble models, which output is a sum of homogeneous sub-mo"},"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":"1905.07325","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"stat.ML","submitted_at":"2019-05-17T15:24:33Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"f3e3330b1d5a95afbf3705aac4a748d7bf9bbde01d69040f2bc973c59101e5ad","abstract_canon_sha256":"39f5f9bb9083604a90d4ad13688484a2bfad73fb08191515fe675e22484ade90"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-17T23:45:55.881212Z","signature_b64":"ZBaNRNw5jw/B884s/FIc5d6BAKZ83ogvolp5WCY6LU7ETDaH4Fs45KhzAXqAT2+SQcuTqJWr05iK+mWNeMYECg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f91a63cddcf694ce9d0244b9dcc3cbb60977231e32c33045b2307bd82fffdd4d","last_reissued_at":"2026-05-17T23:45:55.880595Z","signature_status":"signed_v1","first_computed_at":"2026-05-17T23:45:55.880595Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Lexicographic and Depth-Sensitive Margins in Homogeneous and Non-Homogeneous Deep Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"stat.ML","authors_text":"Daniel Soudry, Jason D. Lee, Mor Shpigel Nacson, Nathan Srebro, Suriya Gunasekar","submitted_at":"2019-05-17T15:24:33Z","abstract_excerpt":"With an eye toward understanding complexity control in deep learning, we study how infinitesimal regularization or gradient descent optimization lead to margin maximizing solutions in both homogeneous and non-homogeneous models, extending previous work that focused on infinitesimal regularization only in homogeneous models. To this end we study the limit of loss minimization with a diverging norm constraint (the \"constrained path\"), relate it to the limit of a \"margin path\" and characterize the resulting solution. For non-homogeneous ensemble models, which output is a sum of homogeneous sub-mo"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1905.07325","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":""},"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":"1905.07325","created_at":"2026-05-17T23:45:55.880706+00:00"},{"alias_kind":"arxiv_version","alias_value":"1905.07325v1","created_at":"2026-05-17T23:45:55.880706+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1905.07325","created_at":"2026-05-17T23:45:55.880706+00:00"},{"alias_kind":"pith_short_12","alias_value":"7ENGHTO462KM","created_at":"2026-05-18T12:33:12.712433+00:00"},{"alias_kind":"pith_short_16","alias_value":"7ENGHTO462KM5HIC","created_at":"2026-05-18T12:33:12.712433+00:00"},{"alias_kind":"pith_short_8","alias_value":"7ENGHTO4","created_at":"2026-05-18T12:33:12.712433+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2401.01335","citing_title":"Self-Play Fine-Tuning Converts Weak Language Models to Strong Language Models","ref_index":244,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/7ENGHTO462KM5HICIS45ZQ6LWY","json":"https://pith.science/pith/7ENGHTO462KM5HICIS45ZQ6LWY.json","graph_json":"https://pith.science/api/pith-number/7ENGHTO462KM5HICIS45ZQ6LWY/graph.json","events_json":"https://pith.science/api/pith-number/7ENGHTO462KM5HICIS45ZQ6LWY/events.json","paper":"https://pith.science/paper/7ENGHTO4"},"agent_actions":{"view_html":"https://pith.science/pith/7ENGHTO462KM5HICIS45ZQ6LWY","download_json":"https://pith.science/pith/7ENGHTO462KM5HICIS45ZQ6LWY.json","view_paper":"https://pith.science/paper/7ENGHTO4","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1905.07325&json=true","fetch_graph":"https://pith.science/api/pith-number/7ENGHTO462KM5HICIS45ZQ6LWY/graph.json","fetch_events":"https://pith.science/api/pith-number/7ENGHTO462KM5HICIS45ZQ6LWY/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/7ENGHTO462KM5HICIS45ZQ6LWY/action/timestamp_anchor","attest_storage":"https://pith.science/pith/7ENGHTO462KM5HICIS45ZQ6LWY/action/storage_attestation","attest_author":"https://pith.science/pith/7ENGHTO462KM5HICIS45ZQ6LWY/action/author_attestation","sign_citation":"https://pith.science/pith/7ENGHTO462KM5HICIS45ZQ6LWY/action/citation_signature","submit_replication":"https://pith.science/pith/7ENGHTO462KM5HICIS45ZQ6LWY/action/replication_record"}},"created_at":"2026-05-17T23:45:55.880706+00:00","updated_at":"2026-05-17T23:45:55.880706+00:00"}