{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:XC2YWXMK7UNRONXY6JTGSMOLUY","short_pith_number":"pith:XC2YWXMK","schema_version":"1.0","canonical_sha256":"b8b58b5d8afd1b1736f8f2666931cba637f3bbdb212951a3156ddd2fb6d9551a","source":{"kind":"arxiv","id":"2502.00885","version":1},"attestation_state":"computed","paper":{"title":"Algorithmic Stability of Stochastic Gradient Descent with Momentum under Heavy-Tailed Noise","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG","math.OC","math.PR"],"primary_cat":"stat.ML","authors_text":"A K M Rokonuzzaman Sonet, Lingjiong Zhu, Melih Barsbey, Mert Gurbuzbalaban, Thanh Dang, Umut Simsekli","submitted_at":"2025-02-02T19:25:48Z","abstract_excerpt":"Understanding the generalization properties of optimization algorithms under heavy-tailed noise has gained growing attention. However, the existing theoretical results mainly focus on stochastic gradient descent (SGD) and the analysis of heavy-tailed optimizers beyond SGD is still missing. In this work, we establish generalization bounds for SGD with momentum (SGDm) under heavy-tailed gradient noise. We first consider the continuous-time limit of SGDm, i.e., a Levy-driven stochastic differential equation (SDE), and establish quantitative Wasserstein algorithmic stability bounds for a class of "},"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":"2502.00885","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"stat.ML","submitted_at":"2025-02-02T19:25:48Z","cross_cats_sorted":["cs.LG","math.OC","math.PR"],"title_canon_sha256":"8d143a72672ccbebfb096b1d8c16426dbafcdaa5da69c3143dd40e94b8b1ebce","abstract_canon_sha256":"2e525b2759e3826f8766a9c6adcb2387f8ee02f29437d9da05b3f02545cb4adf"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:08:39.613548Z","signature_b64":"PJPY9QM9RXgOKG/mOfzvvvDnDtWVmt1H55GHMHdKtlNWRum3ZLThjD6si9Oo/JGKhru0ANFbEv2ctCG7rFcGAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b8b58b5d8afd1b1736f8f2666931cba637f3bbdb212951a3156ddd2fb6d9551a","last_reissued_at":"2026-07-05T10:08:39.613066Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:08:39.613066Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Algorithmic Stability of Stochastic Gradient Descent with Momentum under Heavy-Tailed Noise","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG","math.OC","math.PR"],"primary_cat":"stat.ML","authors_text":"A K M Rokonuzzaman Sonet, Lingjiong Zhu, Melih Barsbey, Mert Gurbuzbalaban, Thanh Dang, Umut Simsekli","submitted_at":"2025-02-02T19:25:48Z","abstract_excerpt":"Understanding the generalization properties of optimization algorithms under heavy-tailed noise has gained growing attention. However, the existing theoretical results mainly focus on stochastic gradient descent (SGD) and the analysis of heavy-tailed optimizers beyond SGD is still missing. In this work, we establish generalization bounds for SGD with momentum (SGDm) under heavy-tailed gradient noise. We first consider the continuous-time limit of SGDm, i.e., a Levy-driven stochastic differential equation (SDE), and establish quantitative Wasserstein algorithmic stability bounds for a class of "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.00885","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/2502.00885/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":"2502.00885","created_at":"2026-07-05T10:08:39.613123+00:00"},{"alias_kind":"arxiv_version","alias_value":"2502.00885v1","created_at":"2026-07-05T10:08:39.613123+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.00885","created_at":"2026-07-05T10:08:39.613123+00:00"},{"alias_kind":"pith_short_12","alias_value":"XC2YWXMK7UNR","created_at":"2026-07-05T10:08:39.613123+00:00"},{"alias_kind":"pith_short_16","alias_value":"XC2YWXMK7UNRONXY","created_at":"2026-07-05T10:08:39.613123+00:00"},{"alias_kind":"pith_short_8","alias_value":"XC2YWXMK","created_at":"2026-07-05T10:08:39.613123+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.28517","citing_title":"Stochastic Gradient Descent with Momentum is Algorithmically Stable","ref_index":12,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/XC2YWXMK7UNRONXY6JTGSMOLUY","json":"https://pith.science/pith/XC2YWXMK7UNRONXY6JTGSMOLUY.json","graph_json":"https://pith.science/api/pith-number/XC2YWXMK7UNRONXY6JTGSMOLUY/graph.json","events_json":"https://pith.science/api/pith-number/XC2YWXMK7UNRONXY6JTGSMOLUY/events.json","paper":"https://pith.science/paper/XC2YWXMK"},"agent_actions":{"view_html":"https://pith.science/pith/XC2YWXMK7UNRONXY6JTGSMOLUY","download_json":"https://pith.science/pith/XC2YWXMK7UNRONXY6JTGSMOLUY.json","view_paper":"https://pith.science/paper/XC2YWXMK","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2502.00885&json=true","fetch_graph":"https://pith.science/api/pith-number/XC2YWXMK7UNRONXY6JTGSMOLUY/graph.json","fetch_events":"https://pith.science/api/pith-number/XC2YWXMK7UNRONXY6JTGSMOLUY/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/XC2YWXMK7UNRONXY6JTGSMOLUY/action/timestamp_anchor","attest_storage":"https://pith.science/pith/XC2YWXMK7UNRONXY6JTGSMOLUY/action/storage_attestation","attest_author":"https://pith.science/pith/XC2YWXMK7UNRONXY6JTGSMOLUY/action/author_attestation","sign_citation":"https://pith.science/pith/XC2YWXMK7UNRONXY6JTGSMOLUY/action/citation_signature","submit_replication":"https://pith.science/pith/XC2YWXMK7UNRONXY6JTGSMOLUY/action/replication_record"}},"created_at":"2026-07-05T10:08:39.613123+00:00","updated_at":"2026-07-05T10:08:39.613123+00:00"}