{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2020:RRGAYXSJRERIBMTUHTMQY33QHC","short_pith_number":"pith:RRGAYXSJ","schema_version":"1.0","canonical_sha256":"8c4c0c5e49892280b2743cd90c6f7038bc64a6722ed2ed6cbc727ed12626f809","source":{"kind":"arxiv","id":"2002.10376","version":1},"attestation_state":"computed","paper":{"title":"The Two Regimes of Deep Network Training","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Aleksander Madry, Guillaume Leclerc","submitted_at":"2020-02-24T17:08:24Z","abstract_excerpt":"Learning rate schedule has a major impact on the performance of deep learning models. Still, the choice of a schedule is often heuristical. We aim to develop a precise understanding of the effects of different learning rate schedules and the appropriate way to select them. To this end, we isolate two distinct phases of training, the first, which we refer to as the \"large-step\" regime, exhibits a rather poor performance from an optimization point of view but is the primary contributor to model generalization; the latter, \"small-step\" regime exhibits much more \"convex-like\" optimization behavior"},"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":"2002.10376","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2020-02-24T17:08:24Z","cross_cats_sorted":["stat.ML"],"title_canon_sha256":"b9012193074021a304e373c0d4cf860a87a7528c557b1f58bcb697c8ade8b585","abstract_canon_sha256":"81cc8de01c2cf9fb1d09d278baee79254bce42751a52fac0514554b7cbed5cee"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T00:43:16.925841Z","signature_b64":"6fKJwA8OZJaPsZSDEL4vjTidIIRfIHXDCtaRKuk6kU6D8CRT+2zQan3VUL1H2AiUzShHpx7Xe733aRl6G3SIAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"8c4c0c5e49892280b2743cd90c6f7038bc64a6722ed2ed6cbc727ed12626f809","last_reissued_at":"2026-07-05T00:43:16.925366Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T00:43:16.925366Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"The Two Regimes of Deep Network Training","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Aleksander Madry, Guillaume Leclerc","submitted_at":"2020-02-24T17:08:24Z","abstract_excerpt":"Learning rate schedule has a major impact on the performance of deep learning models. Still, the choice of a schedule is often heuristical. We aim to develop a precise understanding of the effects of different learning rate schedules and the appropriate way to select them. To this end, we isolate two distinct phases of training, the first, which we refer to as the \"large-step\" regime, exhibits a rather poor performance from an optimization point of view but is the primary contributor to model generalization; the latter, \"small-step\" regime exhibits much more \"convex-like\" optimization behavior"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2002.10376","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/2002.10376/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":"2002.10376","created_at":"2026-07-05T00:43:16.925426+00:00"},{"alias_kind":"arxiv_version","alias_value":"2002.10376v1","created_at":"2026-07-05T00:43:16.925426+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2002.10376","created_at":"2026-07-05T00:43:16.925426+00:00"},{"alias_kind":"pith_short_12","alias_value":"RRGAYXSJRERI","created_at":"2026-07-05T00:43:16.925426+00:00"},{"alias_kind":"pith_short_16","alias_value":"RRGAYXSJRERIBMTU","created_at":"2026-07-05T00:43:16.925426+00:00"},{"alias_kind":"pith_short_8","alias_value":"RRGAYXSJ","created_at":"2026-07-05T00:43:16.925426+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.07495","citing_title":"Second-Order Path Kernel Interpolation Formulas in Machine Learning","ref_index":64,"is_internal_anchor":false},{"citing_arxiv_id":"2603.10079","citing_title":"Large Spikes in Stochastic Gradient Descent: A Large-Deviations View","ref_index":32,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/RRGAYXSJRERIBMTUHTMQY33QHC","json":"https://pith.science/pith/RRGAYXSJRERIBMTUHTMQY33QHC.json","graph_json":"https://pith.science/api/pith-number/RRGAYXSJRERIBMTUHTMQY33QHC/graph.json","events_json":"https://pith.science/api/pith-number/RRGAYXSJRERIBMTUHTMQY33QHC/events.json","paper":"https://pith.science/paper/RRGAYXSJ"},"agent_actions":{"view_html":"https://pith.science/pith/RRGAYXSJRERIBMTUHTMQY33QHC","download_json":"https://pith.science/pith/RRGAYXSJRERIBMTUHTMQY33QHC.json","view_paper":"https://pith.science/paper/RRGAYXSJ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2002.10376&json=true","fetch_graph":"https://pith.science/api/pith-number/RRGAYXSJRERIBMTUHTMQY33QHC/graph.json","fetch_events":"https://pith.science/api/pith-number/RRGAYXSJRERIBMTUHTMQY33QHC/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/RRGAYXSJRERIBMTUHTMQY33QHC/action/timestamp_anchor","attest_storage":"https://pith.science/pith/RRGAYXSJRERIBMTUHTMQY33QHC/action/storage_attestation","attest_author":"https://pith.science/pith/RRGAYXSJRERIBMTUHTMQY33QHC/action/author_attestation","sign_citation":"https://pith.science/pith/RRGAYXSJRERIBMTUHTMQY33QHC/action/citation_signature","submit_replication":"https://pith.science/pith/RRGAYXSJRERIBMTUHTMQY33QHC/action/replication_record"}},"created_at":"2026-07-05T00:43:16.925426+00:00","updated_at":"2026-07-05T00:43:16.925426+00:00"}