{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2020:NGM4BDAXCVBK5PF4W3NV2L4SSI","short_pith_number":"pith:NGM4BDAX","schema_version":"1.0","canonical_sha256":"6999c08c171542aebcbcb6db5d2f929237b125f99aba14dc3731e57a00035bd9","source":{"kind":"arxiv","id":"2006.06293","version":1},"attestation_state":"computed","paper":{"title":"Multiplicative noise and heavy tails in stochastic optimization","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG","math.OC","math.ST","stat.TH"],"primary_cat":"stat.ML","authors_text":"Liam Hodgkinson, Michael W. Mahoney","submitted_at":"2020-06-11T09:58:01Z","abstract_excerpt":"Although stochastic optimization is central to modern machine learning, the precise mechanisms underlying its success, and in particular, the precise role of the stochasticity, still remain unclear. Modelling stochastic optimization algorithms as discrete random recurrence relations, we show that multiplicative noise, as it commonly arises due to variance in local rates of convergence, results in heavy-tailed stationary behaviour in the parameters. A detailed analysis is conducted for SGD applied to a simple linear regression problem, followed by theoretical results for a much larger 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":"2006.06293","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"stat.ML","submitted_at":"2020-06-11T09:58:01Z","cross_cats_sorted":["cs.LG","math.OC","math.ST","stat.TH"],"title_canon_sha256":"e37486110dc613a625a58b9956eba31e19624846c18191637663f576d44e43e6","abstract_canon_sha256":"55a9f75076e222a7d74c4e2e4813e3de3ad0e08000c1682d6f8f64e9bd1093f6"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T01:09:32.470076Z","signature_b64":"Um4UtLPx89OOyxdvHmCV3mJzT+85XicQQcB+3wr7N98EmjUWzEfp+s3bhYXFCDm2aoE/KAtM4NlCAI6OJYECDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"6999c08c171542aebcbcb6db5d2f929237b125f99aba14dc3731e57a00035bd9","last_reissued_at":"2026-07-05T01:09:32.469602Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T01:09:32.469602Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Multiplicative noise and heavy tails in stochastic optimization","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG","math.OC","math.ST","stat.TH"],"primary_cat":"stat.ML","authors_text":"Liam Hodgkinson, Michael W. Mahoney","submitted_at":"2020-06-11T09:58:01Z","abstract_excerpt":"Although stochastic optimization is central to modern machine learning, the precise mechanisms underlying its success, and in particular, the precise role of the stochasticity, still remain unclear. Modelling stochastic optimization algorithms as discrete random recurrence relations, we show that multiplicative noise, as it commonly arises due to variance in local rates of convergence, results in heavy-tailed stationary behaviour in the parameters. A detailed analysis is conducted for SGD applied to a simple linear regression problem, followed by theoretical results for a much larger class of "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2006.06293","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/2006.06293/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":"2006.06293","created_at":"2026-07-05T01:09:32.469662+00:00"},{"alias_kind":"arxiv_version","alias_value":"2006.06293v1","created_at":"2026-07-05T01:09:32.469662+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2006.06293","created_at":"2026-07-05T01:09:32.469662+00:00"},{"alias_kind":"pith_short_12","alias_value":"NGM4BDAXCVBK","created_at":"2026-07-05T01:09:32.469662+00:00"},{"alias_kind":"pith_short_16","alias_value":"NGM4BDAXCVBK5PF4","created_at":"2026-07-05T01:09:32.469662+00:00"},{"alias_kind":"pith_short_8","alias_value":"NGM4BDAX","created_at":"2026-07-05T01:09:32.469662+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2603.09355","citing_title":"SHANG++: Robust Stochastic Acceleration under Multiplicative Noise","ref_index":5,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/NGM4BDAXCVBK5PF4W3NV2L4SSI","json":"https://pith.science/pith/NGM4BDAXCVBK5PF4W3NV2L4SSI.json","graph_json":"https://pith.science/api/pith-number/NGM4BDAXCVBK5PF4W3NV2L4SSI/graph.json","events_json":"https://pith.science/api/pith-number/NGM4BDAXCVBK5PF4W3NV2L4SSI/events.json","paper":"https://pith.science/paper/NGM4BDAX"},"agent_actions":{"view_html":"https://pith.science/pith/NGM4BDAXCVBK5PF4W3NV2L4SSI","download_json":"https://pith.science/pith/NGM4BDAXCVBK5PF4W3NV2L4SSI.json","view_paper":"https://pith.science/paper/NGM4BDAX","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2006.06293&json=true","fetch_graph":"https://pith.science/api/pith-number/NGM4BDAXCVBK5PF4W3NV2L4SSI/graph.json","fetch_events":"https://pith.science/api/pith-number/NGM4BDAXCVBK5PF4W3NV2L4SSI/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/NGM4BDAXCVBK5PF4W3NV2L4SSI/action/timestamp_anchor","attest_storage":"https://pith.science/pith/NGM4BDAXCVBK5PF4W3NV2L4SSI/action/storage_attestation","attest_author":"https://pith.science/pith/NGM4BDAXCVBK5PF4W3NV2L4SSI/action/author_attestation","sign_citation":"https://pith.science/pith/NGM4BDAXCVBK5PF4W3NV2L4SSI/action/citation_signature","submit_replication":"https://pith.science/pith/NGM4BDAXCVBK5PF4W3NV2L4SSI/action/replication_record"}},"created_at":"2026-07-05T01:09:32.469662+00:00","updated_at":"2026-07-05T01:09:32.469662+00:00"}