{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:3722GYFVSYAP34LDU7Q5MVOE4L","short_pith_number":"pith:3722GYFV","schema_version":"1.0","canonical_sha256":"dff5a360b59600fdf163a7e1d655c4e2ca27c223a3b7906f3f74565ca6bfa578","source":{"kind":"arxiv","id":"2302.10181","version":1},"attestation_state":"computed","paper":{"title":"Exploring the Effect of Multi-step Ascent in Sharpness-Aware Minimization","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Hoki Kim, Jaewook Lee, Jinseong Park, Woojin Lee, Yujin Choi","submitted_at":"2023-01-27T06:16:31Z","abstract_excerpt":"Recently, Sharpness-Aware Minimization (SAM) has shown state-of-the-art performance by seeking flat minima. To minimize the maximum loss within a neighborhood in the parameter space, SAM uses an ascent step, which perturbs the weights along the direction of gradient ascent with a given radius. While single-step or multi-step can be taken during ascent steps, previous studies have shown that multi-step ascent SAM rarely improves generalization performance. However, this phenomenon is particularly interesting because the multi-step ascent is expected to provide a better approximation of the maxi"},"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":"2302.10181","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.LG","submitted_at":"2023-01-27T06:16:31Z","cross_cats_sorted":[],"title_canon_sha256":"0f2930b20bf634453384c1a026548226b5b4c238b0540eb6b2e4993bdc6c4ede","abstract_canon_sha256":"0165496c5c1be25bc29abb24b50715453f90d01129c103618186f50e3d96af13"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T05:43:36.208285Z","signature_b64":"cTBeycNHrIkzywey+uSBZ4vsRwyaU0pfm0F4aWIBi4oNv+LW302TGksbXaPtkegThi0qXchZ+6hT9rDO2x5dDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"dff5a360b59600fdf163a7e1d655c4e2ca27c223a3b7906f3f74565ca6bfa578","last_reissued_at":"2026-07-05T05:43:36.207796Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T05:43:36.207796Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Exploring the Effect of Multi-step Ascent in Sharpness-Aware Minimization","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Hoki Kim, Jaewook Lee, Jinseong Park, Woojin Lee, Yujin Choi","submitted_at":"2023-01-27T06:16:31Z","abstract_excerpt":"Recently, Sharpness-Aware Minimization (SAM) has shown state-of-the-art performance by seeking flat minima. To minimize the maximum loss within a neighborhood in the parameter space, SAM uses an ascent step, which perturbs the weights along the direction of gradient ascent with a given radius. While single-step or multi-step can be taken during ascent steps, previous studies have shown that multi-step ascent SAM rarely improves generalization performance. However, this phenomenon is particularly interesting because the multi-step ascent is expected to provide a better approximation of the maxi"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2302.10181","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/2302.10181/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":"2302.10181","created_at":"2026-07-05T05:43:36.207871+00:00"},{"alias_kind":"arxiv_version","alias_value":"2302.10181v1","created_at":"2026-07-05T05:43:36.207871+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2302.10181","created_at":"2026-07-05T05:43:36.207871+00:00"},{"alias_kind":"pith_short_12","alias_value":"3722GYFVSYAP","created_at":"2026-07-05T05:43:36.207871+00:00"},{"alias_kind":"pith_short_16","alias_value":"3722GYFVSYAP34LD","created_at":"2026-07-05T05:43:36.207871+00:00"},{"alias_kind":"pith_short_8","alias_value":"3722GYFV","created_at":"2026-07-05T05:43:36.207871+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2509.02433","citing_title":"VASSO: Variance Suppression for Sharpness-Aware Minimization","ref_index":60,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/3722GYFVSYAP34LDU7Q5MVOE4L","json":"https://pith.science/pith/3722GYFVSYAP34LDU7Q5MVOE4L.json","graph_json":"https://pith.science/api/pith-number/3722GYFVSYAP34LDU7Q5MVOE4L/graph.json","events_json":"https://pith.science/api/pith-number/3722GYFVSYAP34LDU7Q5MVOE4L/events.json","paper":"https://pith.science/paper/3722GYFV"},"agent_actions":{"view_html":"https://pith.science/pith/3722GYFVSYAP34LDU7Q5MVOE4L","download_json":"https://pith.science/pith/3722GYFVSYAP34LDU7Q5MVOE4L.json","view_paper":"https://pith.science/paper/3722GYFV","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2302.10181&json=true","fetch_graph":"https://pith.science/api/pith-number/3722GYFVSYAP34LDU7Q5MVOE4L/graph.json","fetch_events":"https://pith.science/api/pith-number/3722GYFVSYAP34LDU7Q5MVOE4L/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/3722GYFVSYAP34LDU7Q5MVOE4L/action/timestamp_anchor","attest_storage":"https://pith.science/pith/3722GYFVSYAP34LDU7Q5MVOE4L/action/storage_attestation","attest_author":"https://pith.science/pith/3722GYFVSYAP34LDU7Q5MVOE4L/action/author_attestation","sign_citation":"https://pith.science/pith/3722GYFVSYAP34LDU7Q5MVOE4L/action/citation_signature","submit_replication":"https://pith.science/pith/3722GYFVSYAP34LDU7Q5MVOE4L/action/replication_record"}},"created_at":"2026-07-05T05:43:36.207871+00:00","updated_at":"2026-07-05T05:43:36.207871+00:00"}