{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2020:7CUSZC4KXGJVHYMPYSCJOOZ3KB","short_pith_number":"pith:7CUSZC4K","schema_version":"1.0","canonical_sha256":"f8a92c8b8ab99353e18fc484973b3b507cb4b286fcca12baac8c6476ee446a6c","source":{"kind":"arxiv","id":"2006.11226","version":1},"attestation_state":"computed","paper":{"title":"Gradient descent follows the regularization path for general losses","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["math.OC","stat.ML"],"primary_cat":"cs.LG","authors_text":"Matus Telgarsky, Miroslav Dud\\'ik, Robert E. Schapire, Ziwei Ji","submitted_at":"2020-06-19T17:01:25Z","abstract_excerpt":"Recent work across many machine learning disciplines has highlighted that standard descent methods, even without explicit regularization, do not merely minimize the training error, but also exhibit an implicit bias. This bias is typically towards a certain regularized solution, and relies upon the details of the learning process, for instance the use of the cross-entropy loss.\n  In this work, we show that for empirical risk minimization over linear predictors with arbitrary convex, strictly decreasing losses, if the risk does not attain its infimum, then the gradient-descent path and the algor"},"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":"2006.11226","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2020-06-19T17:01:25Z","cross_cats_sorted":["math.OC","stat.ML"],"title_canon_sha256":"b67102d7a6846c2ecdae44f8b4c1f11d37240d0002e6ce95bbe23cba1201e272","abstract_canon_sha256":"94a2a42113f49301295bba6772961fdb2f7c19933464ccb48ff35c445902d434"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T01:11:36.638903Z","signature_b64":"ySJCCRvvPuEYJnSsTAYlcjfkIItJ8wplhKgBYBzezyphhlL0nDVjwkTu4wavTBWk+CbgDhPFKGl5jpSQ83uMAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f8a92c8b8ab99353e18fc484973b3b507cb4b286fcca12baac8c6476ee446a6c","last_reissued_at":"2026-07-05T01:11:36.638498Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T01:11:36.638498Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Gradient descent follows the regularization path for general losses","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["math.OC","stat.ML"],"primary_cat":"cs.LG","authors_text":"Matus Telgarsky, Miroslav Dud\\'ik, Robert E. Schapire, Ziwei Ji","submitted_at":"2020-06-19T17:01:25Z","abstract_excerpt":"Recent work across many machine learning disciplines has highlighted that standard descent methods, even without explicit regularization, do not merely minimize the training error, but also exhibit an implicit bias. This bias is typically towards a certain regularized solution, and relies upon the details of the learning process, for instance the use of the cross-entropy loss.\n  In this work, we show that for empirical risk minimization over linear predictors with arbitrary convex, strictly decreasing losses, if the risk does not attain its infimum, then the gradient-descent path and the algor"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2006.11226","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/2006.11226/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":"2006.11226","created_at":"2026-07-05T01:11:36.638554+00:00"},{"alias_kind":"arxiv_version","alias_value":"2006.11226v1","created_at":"2026-07-05T01:11:36.638554+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2006.11226","created_at":"2026-07-05T01:11:36.638554+00:00"},{"alias_kind":"pith_short_12","alias_value":"7CUSZC4KXGJV","created_at":"2026-07-05T01:11:36.638554+00:00"},{"alias_kind":"pith_short_16","alias_value":"7CUSZC4KXGJVHYMP","created_at":"2026-07-05T01:11:36.638554+00:00"},{"alias_kind":"pith_short_8","alias_value":"7CUSZC4K","created_at":"2026-07-05T01:11:36.638554+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/7CUSZC4KXGJVHYMPYSCJOOZ3KB","json":"https://pith.science/pith/7CUSZC4KXGJVHYMPYSCJOOZ3KB.json","graph_json":"https://pith.science/api/pith-number/7CUSZC4KXGJVHYMPYSCJOOZ3KB/graph.json","events_json":"https://pith.science/api/pith-number/7CUSZC4KXGJVHYMPYSCJOOZ3KB/events.json","paper":"https://pith.science/paper/7CUSZC4K"},"agent_actions":{"view_html":"https://pith.science/pith/7CUSZC4KXGJVHYMPYSCJOOZ3KB","download_json":"https://pith.science/pith/7CUSZC4KXGJVHYMPYSCJOOZ3KB.json","view_paper":"https://pith.science/paper/7CUSZC4K","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2006.11226&json=true","fetch_graph":"https://pith.science/api/pith-number/7CUSZC4KXGJVHYMPYSCJOOZ3KB/graph.json","fetch_events":"https://pith.science/api/pith-number/7CUSZC4KXGJVHYMPYSCJOOZ3KB/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/7CUSZC4KXGJVHYMPYSCJOOZ3KB/action/timestamp_anchor","attest_storage":"https://pith.science/pith/7CUSZC4KXGJVHYMPYSCJOOZ3KB/action/storage_attestation","attest_author":"https://pith.science/pith/7CUSZC4KXGJVHYMPYSCJOOZ3KB/action/author_attestation","sign_citation":"https://pith.science/pith/7CUSZC4KXGJVHYMPYSCJOOZ3KB/action/citation_signature","submit_replication":"https://pith.science/pith/7CUSZC4KXGJVHYMPYSCJOOZ3KB/action/replication_record"}},"created_at":"2026-07-05T01:11:36.638554+00:00","updated_at":"2026-07-05T01:11:36.638554+00:00"}