{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:X76YEN3MXQCQ6ASSUEIZ7F2R4K","short_pith_number":"pith:X76YEN3M","schema_version":"1.0","canonical_sha256":"bffd82376cbc050f0252a1119f9751e28eb984a9bf97feeb085827eeff84a04f","source":{"kind":"arxiv","id":"2302.00849","version":1},"attestation_state":"computed","paper":{"title":"Implicit regularization in Heavy-ball momentum accelerated stochastic gradient descent","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["math.OC"],"primary_cat":"cs.LG","authors_text":"Avrajit Ghosh, He Lyu, Rongrong Wang, Xitong Zhang","submitted_at":"2023-02-02T03:25:38Z","abstract_excerpt":"It is well known that the finite step-size ($h$) in Gradient Descent (GD) implicitly regularizes solutions to flatter minima. A natural question to ask is \"Does the momentum parameter $\\beta$ play a role in implicit regularization in Heavy-ball (H.B) momentum accelerated gradient descent (GD+M)?\". To answer this question, first, we show that the discrete H.B momentum update (GD+M) follows a continuous trajectory induced by a modified loss, which consists of an original loss and an implicit regularizer. Then, we show that this implicit regularizer for (GD+M) is stronger than that of (GD) by fac"},"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":"2302.00849","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2023-02-02T03:25:38Z","cross_cats_sorted":["math.OC"],"title_canon_sha256":"e2d5fb5f6d754d1b1c5d37deffddae5424cd1358e02938419a8cd3c4b926c6e5","abstract_canon_sha256":"3b6fe8d248c1f4e009586cd94830c01dff49741627ac6fc120886abf9a3092ed"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T05:38:13.108370Z","signature_b64":"iEia68hQrtkTfDE9UsoojS1HQ4VgQsHdNZn7XLRDxsmp8lVnPYLALSd/OnHaB58R0roCG3ESybK8L9JHx+u7Bg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"bffd82376cbc050f0252a1119f9751e28eb984a9bf97feeb085827eeff84a04f","last_reissued_at":"2026-07-05T05:38:13.107902Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T05:38:13.107902Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Implicit regularization in Heavy-ball momentum accelerated stochastic gradient descent","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["math.OC"],"primary_cat":"cs.LG","authors_text":"Avrajit Ghosh, He Lyu, Rongrong Wang, Xitong Zhang","submitted_at":"2023-02-02T03:25:38Z","abstract_excerpt":"It is well known that the finite step-size ($h$) in Gradient Descent (GD) implicitly regularizes solutions to flatter minima. A natural question to ask is \"Does the momentum parameter $\\beta$ play a role in implicit regularization in Heavy-ball (H.B) momentum accelerated gradient descent (GD+M)?\". To answer this question, first, we show that the discrete H.B momentum update (GD+M) follows a continuous trajectory induced by a modified loss, which consists of an original loss and an implicit regularizer. Then, we show that this implicit regularizer for (GD+M) is stronger than that of (GD) by fac"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2302.00849","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/2302.00849/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":"2302.00849","created_at":"2026-07-05T05:38:13.107960+00:00"},{"alias_kind":"arxiv_version","alias_value":"2302.00849v1","created_at":"2026-07-05T05:38:13.107960+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2302.00849","created_at":"2026-07-05T05:38:13.107960+00:00"},{"alias_kind":"pith_short_12","alias_value":"X76YEN3MXQCQ","created_at":"2026-07-05T05:38:13.107960+00:00"},{"alias_kind":"pith_short_16","alias_value":"X76YEN3MXQCQ6ASS","created_at":"2026-07-05T05:38:13.107960+00:00"},{"alias_kind":"pith_short_8","alias_value":"X76YEN3M","created_at":"2026-07-05T05:38:13.107960+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.29152","citing_title":"Do Deep Networks Forget Initialization? A Forgetting-Time View of Practical Inductive Bias","ref_index":39,"is_internal_anchor":false},{"citing_arxiv_id":"2605.14345","citing_title":"Convergence of difference inclusions via a diameter criterion","ref_index":89,"is_internal_anchor":false},{"citing_arxiv_id":"2604.14108","citing_title":"Momentum Further Constrains Sharpness at the Edge of Stochastic Stability","ref_index":12,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/X76YEN3MXQCQ6ASSUEIZ7F2R4K","json":"https://pith.science/pith/X76YEN3MXQCQ6ASSUEIZ7F2R4K.json","graph_json":"https://pith.science/api/pith-number/X76YEN3MXQCQ6ASSUEIZ7F2R4K/graph.json","events_json":"https://pith.science/api/pith-number/X76YEN3MXQCQ6ASSUEIZ7F2R4K/events.json","paper":"https://pith.science/paper/X76YEN3M"},"agent_actions":{"view_html":"https://pith.science/pith/X76YEN3MXQCQ6ASSUEIZ7F2R4K","download_json":"https://pith.science/pith/X76YEN3MXQCQ6ASSUEIZ7F2R4K.json","view_paper":"https://pith.science/paper/X76YEN3M","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2302.00849&json=true","fetch_graph":"https://pith.science/api/pith-number/X76YEN3MXQCQ6ASSUEIZ7F2R4K/graph.json","fetch_events":"https://pith.science/api/pith-number/X76YEN3MXQCQ6ASSUEIZ7F2R4K/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/X76YEN3MXQCQ6ASSUEIZ7F2R4K/action/timestamp_anchor","attest_storage":"https://pith.science/pith/X76YEN3MXQCQ6ASSUEIZ7F2R4K/action/storage_attestation","attest_author":"https://pith.science/pith/X76YEN3MXQCQ6ASSUEIZ7F2R4K/action/author_attestation","sign_citation":"https://pith.science/pith/X76YEN3MXQCQ6ASSUEIZ7F2R4K/action/citation_signature","submit_replication":"https://pith.science/pith/X76YEN3MXQCQ6ASSUEIZ7F2R4K/action/replication_record"}},"created_at":"2026-07-05T05:38:13.107960+00:00","updated_at":"2026-07-05T05:38:13.107960+00:00"}