{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2019:P7F6K57TCAXRAZ53DWPOQ24G2V","short_pith_number":"pith:P7F6K57T","schema_version":"1.0","canonical_sha256":"7fcbe577f3102f1067bb1d9ee86b86d550a4a251622238dc58b4f02950c2281b","source":{"kind":"arxiv","id":"1910.07454","version":3},"attestation_state":"computed","paper":{"title":"An Exponential Learning Rate Schedule for Deep Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Sanjeev Arora, Zhiyuan Li","submitted_at":"2019-10-16T16:22:58Z","abstract_excerpt":"Intriguing empirical evidence exists that deep learning can work well with exoticschedules for varying the learning rate. This paper suggests that the phenomenon may be due to Batch Normalization or BN, which is ubiquitous and provides benefits in optimization and generalization across all standard architectures. The following new results are shown about BN with weight decay and momentum (in other words, the typical use case which was not considered in earlier theoretical analyses of stand-alone BN.\n  1. Training can be done using SGD with momentum and an exponentially increasing learning rate"},"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":"1910.07454","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2019-10-16T16:22:58Z","cross_cats_sorted":["stat.ML"],"title_canon_sha256":"5532ff4894f57a00ed766aa901aed7d9238b64a1629d0b7f0023735d0aac02ab","abstract_canon_sha256":"db00ef397c051e1d649810009f387730a66453e2e53b0922d1d4a5b672942cfa"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T00:20:55.832889Z","signature_b64":"BkCn8Koq4gM8CrGQ2NZYbRS1wjQiTiEMlj1OqjB6wRsqr83qkxrNoaY948eYIlaHOKAH7VEVsM5FZbF/Q06qBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"7fcbe577f3102f1067bb1d9ee86b86d550a4a251622238dc58b4f02950c2281b","last_reissued_at":"2026-07-05T00:20:55.832433Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T00:20:55.832433Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"An Exponential Learning Rate Schedule for Deep Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Sanjeev Arora, Zhiyuan Li","submitted_at":"2019-10-16T16:22:58Z","abstract_excerpt":"Intriguing empirical evidence exists that deep learning can work well with exoticschedules for varying the learning rate. This paper suggests that the phenomenon may be due to Batch Normalization or BN, which is ubiquitous and provides benefits in optimization and generalization across all standard architectures. The following new results are shown about BN with weight decay and momentum (in other words, the typical use case which was not considered in earlier theoretical analyses of stand-alone BN.\n  1. Training can be done using SGD with momentum and an exponentially increasing learning rate"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1910.07454","kind":"arxiv","version":3},"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/1910.07454/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":"1910.07454","created_at":"2026-07-05T00:20:55.832490+00:00"},{"alias_kind":"arxiv_version","alias_value":"1910.07454v3","created_at":"2026-07-05T00:20:55.832490+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1910.07454","created_at":"2026-07-05T00:20:55.832490+00:00"},{"alias_kind":"pith_short_12","alias_value":"P7F6K57TCAXR","created_at":"2026-07-05T00:20:55.832490+00:00"},{"alias_kind":"pith_short_16","alias_value":"P7F6K57TCAXRAZ53","created_at":"2026-07-05T00:20:55.832490+00:00"},{"alias_kind":"pith_short_8","alias_value":"P7F6K57T","created_at":"2026-07-05T00:20:55.832490+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.25971","citing_title":"Improving Neural Network Training by Decoupling the Magnitude and Direction of Weight Vectors","ref_index":162,"is_internal_anchor":false},{"citing_arxiv_id":"2606.19302","citing_title":"Optimal scenario design for climate emulation","ref_index":235,"is_internal_anchor":false},{"citing_arxiv_id":"2507.15970","citing_title":"CIS-BWE: Chaos-Informed Speech Bandwidth Extension","ref_index":6,"is_internal_anchor":false},{"citing_arxiv_id":"2605.04418","citing_title":"Demystifying Manifold Constraints in LLM Pre-training","ref_index":58,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/P7F6K57TCAXRAZ53DWPOQ24G2V","json":"https://pith.science/pith/P7F6K57TCAXRAZ53DWPOQ24G2V.json","graph_json":"https://pith.science/api/pith-number/P7F6K57TCAXRAZ53DWPOQ24G2V/graph.json","events_json":"https://pith.science/api/pith-number/P7F6K57TCAXRAZ53DWPOQ24G2V/events.json","paper":"https://pith.science/paper/P7F6K57T"},"agent_actions":{"view_html":"https://pith.science/pith/P7F6K57TCAXRAZ53DWPOQ24G2V","download_json":"https://pith.science/pith/P7F6K57TCAXRAZ53DWPOQ24G2V.json","view_paper":"https://pith.science/paper/P7F6K57T","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1910.07454&json=true","fetch_graph":"https://pith.science/api/pith-number/P7F6K57TCAXRAZ53DWPOQ24G2V/graph.json","fetch_events":"https://pith.science/api/pith-number/P7F6K57TCAXRAZ53DWPOQ24G2V/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/P7F6K57TCAXRAZ53DWPOQ24G2V/action/timestamp_anchor","attest_storage":"https://pith.science/pith/P7F6K57TCAXRAZ53DWPOQ24G2V/action/storage_attestation","attest_author":"https://pith.science/pith/P7F6K57TCAXRAZ53DWPOQ24G2V/action/author_attestation","sign_citation":"https://pith.science/pith/P7F6K57TCAXRAZ53DWPOQ24G2V/action/citation_signature","submit_replication":"https://pith.science/pith/P7F6K57TCAXRAZ53DWPOQ24G2V/action/replication_record"}},"created_at":"2026-07-05T00:20:55.832490+00:00","updated_at":"2026-07-05T00:20:55.832490+00:00"}