{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2020:NRMTXU6E3B2R2BGGBSVH3RLTHK","short_pith_number":"pith:NRMTXU6E","schema_version":"1.0","canonical_sha256":"6c593bd3c4d8751d04c60caa7dc5733a8f23169c77c3681571cb13c49b32b289","source":{"kind":"arxiv","id":"2002.11803","version":1},"attestation_state":"computed","paper":{"title":"Disentangling Adaptive Gradient Methods from Learning Rates","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Cyril Zhang, Elad Hazan, Naman Agarwal, Rohan Anil, Tomer Koren","submitted_at":"2020-02-26T21:42:49Z","abstract_excerpt":"We investigate several confounding factors in the evaluation of optimization algorithms for deep learning. Primarily, we take a deeper look at how adaptive gradient methods interact with the learning rate schedule, a notoriously difficult-to-tune hyperparameter which has dramatic effects on the convergence and generalization of neural network training. We introduce a \"grafting\" experiment which decouples an update's magnitude from its direction, finding that many existing beliefs in the literature may have arisen from insufficient isolation of the implicit schedule of step sizes. Alongside thi"},"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":"2002.11803","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2020-02-26T21:42:49Z","cross_cats_sorted":["stat.ML"],"title_canon_sha256":"e2304e3fc35c995f32b5cba948a4c36bfa04fff34bdca500690394bb45ae5b84","abstract_canon_sha256":"b9679d76b890db41d4fe331914349691d71ca2a7d6f6326522de8c5417bc1625"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T00:44:09.160167Z","signature_b64":"RFnVezX2YMQphccB4TjGmXI106GVYyY44LC8TgEFwBKIHSXKSAaykStCVAOo50SPZVbO9VqIZnA3jsQ1HhS4Dg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"6c593bd3c4d8751d04c60caa7dc5733a8f23169c77c3681571cb13c49b32b289","last_reissued_at":"2026-07-05T00:44:09.159735Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T00:44:09.159735Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Disentangling Adaptive Gradient Methods from Learning Rates","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Cyril Zhang, Elad Hazan, Naman Agarwal, Rohan Anil, Tomer Koren","submitted_at":"2020-02-26T21:42:49Z","abstract_excerpt":"We investigate several confounding factors in the evaluation of optimization algorithms for deep learning. Primarily, we take a deeper look at how adaptive gradient methods interact with the learning rate schedule, a notoriously difficult-to-tune hyperparameter which has dramatic effects on the convergence and generalization of neural network training. We introduce a \"grafting\" experiment which decouples an update's magnitude from its direction, finding that many existing beliefs in the literature may have arisen from insufficient isolation of the implicit schedule of step sizes. Alongside thi"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2002.11803","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/2002.11803/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":"2002.11803","created_at":"2026-07-05T00:44:09.159792+00:00"},{"alias_kind":"arxiv_version","alias_value":"2002.11803v1","created_at":"2026-07-05T00:44:09.159792+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2002.11803","created_at":"2026-07-05T00:44:09.159792+00:00"},{"alias_kind":"pith_short_12","alias_value":"NRMTXU6E3B2R","created_at":"2026-07-05T00:44:09.159792+00:00"},{"alias_kind":"pith_short_16","alias_value":"NRMTXU6E3B2R2BGG","created_at":"2026-07-05T00:44:09.159792+00:00"},{"alias_kind":"pith_short_8","alias_value":"NRMTXU6E","created_at":"2026-07-05T00:44:09.159792+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.04058","citing_title":"Spectral Scaling Laws of Muon","ref_index":1,"is_internal_anchor":false},{"citing_arxiv_id":"2606.02365","citing_title":"FOAM: Frequency and Operator Error-Based Adaptive Damping Method for Reducing Staleness-Oriented Error for Shampoo","ref_index":4,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/NRMTXU6E3B2R2BGGBSVH3RLTHK","json":"https://pith.science/pith/NRMTXU6E3B2R2BGGBSVH3RLTHK.json","graph_json":"https://pith.science/api/pith-number/NRMTXU6E3B2R2BGGBSVH3RLTHK/graph.json","events_json":"https://pith.science/api/pith-number/NRMTXU6E3B2R2BGGBSVH3RLTHK/events.json","paper":"https://pith.science/paper/NRMTXU6E"},"agent_actions":{"view_html":"https://pith.science/pith/NRMTXU6E3B2R2BGGBSVH3RLTHK","download_json":"https://pith.science/pith/NRMTXU6E3B2R2BGGBSVH3RLTHK.json","view_paper":"https://pith.science/paper/NRMTXU6E","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2002.11803&json=true","fetch_graph":"https://pith.science/api/pith-number/NRMTXU6E3B2R2BGGBSVH3RLTHK/graph.json","fetch_events":"https://pith.science/api/pith-number/NRMTXU6E3B2R2BGGBSVH3RLTHK/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/NRMTXU6E3B2R2BGGBSVH3RLTHK/action/timestamp_anchor","attest_storage":"https://pith.science/pith/NRMTXU6E3B2R2BGGBSVH3RLTHK/action/storage_attestation","attest_author":"https://pith.science/pith/NRMTXU6E3B2R2BGGBSVH3RLTHK/action/author_attestation","sign_citation":"https://pith.science/pith/NRMTXU6E3B2R2BGGBSVH3RLTHK/action/citation_signature","submit_replication":"https://pith.science/pith/NRMTXU6E3B2R2BGGBSVH3RLTHK/action/replication_record"}},"created_at":"2026-07-05T00:44:09.159792+00:00","updated_at":"2026-07-05T00:44:09.159792+00:00"}