{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:YFBVAKCDZEDEWA53FFYKOUN4VR","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"4401318839c49cc2ae1a4c7be2ef669c4bbcd863f1ab6ce6e7ae8a388ad39ec3","cross_cats_sorted":["cs.AI","stat.ML"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-06-03T17:53:25Z","title_canon_sha256":"4a3110d111fe819e514d8b27789318969f421db63f03d9e17d66305b6147b42c"},"schema_version":"1.0","source":{"id":"2406.01660","kind":"arxiv","version":4}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2406.01660","created_at":"2026-07-05T10:47:54Z"},{"alias_kind":"arxiv_version","alias_value":"2406.01660v4","created_at":"2026-07-05T10:47:54Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2406.01660","created_at":"2026-07-05T10:47:54Z"},{"alias_kind":"pith_short_12","alias_value":"YFBVAKCDZEDE","created_at":"2026-07-05T10:47:54Z"},{"alias_kind":"pith_short_16","alias_value":"YFBVAKCDZEDEWA53","created_at":"2026-07-05T10:47:54Z"},{"alias_kind":"pith_short_8","alias_value":"YFBVAKCD","created_at":"2026-07-05T10:47:54Z"}],"graph_snapshots":[{"event_id":"sha256:0d5f8bc316daffdf7362029e91b9784711a351d5d9e9f068d23cc30147ca7a47","target":"graph","created_at":"2026-07-05T10:47:54Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2406.01660/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Online and offline RLHF methods, such as PPO and DPO, have been highly successful in aligning AI with human preferences. Despite their success, however, these methods suffer from fundamental limitations: (a) Models trained with RLHF can learn from mistakes or negative examples through RL mechanism or contrastive loss during training. However, at inference time, they lack an innate self-improvement mechanism for error corrections. (b) The optimal solution of existing methods is highly task-dependent, making it difficult for them to generalize to new tasks. To address these challenges, we propos","authors_text":"Arash Ahmadian, Eugene Choi, Matthieu Geist, Mohammad Gheshlaghi Azar, Oilvier Pietquin","cross_cats":["cs.AI","stat.ML"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-06-03T17:53:25Z","title":"Self-Improving Robust Preference Optimization"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2406.01660","kind":"arxiv","version":4},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:14ef336e475d443472a4036e8cabc2511f95e2f8137a11395ae2e4ac20f9702a","target":"record","created_at":"2026-07-05T10:47:54Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"4401318839c49cc2ae1a4c7be2ef669c4bbcd863f1ab6ce6e7ae8a388ad39ec3","cross_cats_sorted":["cs.AI","stat.ML"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-06-03T17:53:25Z","title_canon_sha256":"4a3110d111fe819e514d8b27789318969f421db63f03d9e17d66305b6147b42c"},"schema_version":"1.0","source":{"id":"2406.01660","kind":"arxiv","version":4}},"canonical_sha256":"c143502843c9064b03bb2970a751bcac6a9b97224e1c1859dab991a1ff11a99c","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"c143502843c9064b03bb2970a751bcac6a9b97224e1c1859dab991a1ff11a99c","first_computed_at":"2026-07-05T10:47:54.606500Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T10:47:54.606500Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"165zQslMw5Ntot9U8z93eRY6N+pODpwMh04ZatxmvO3gOVdKyAzNROESiWiahOkLt7hw5fo6p2ZoFllsgjwaAw==","signature_status":"signed_v1","signed_at":"2026-07-05T10:47:54.606982Z","signed_message":"canonical_sha256_bytes"},"source_id":"2406.01660","source_kind":"arxiv","source_version":4}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:14ef336e475d443472a4036e8cabc2511f95e2f8137a11395ae2e4ac20f9702a","sha256:0d5f8bc316daffdf7362029e91b9784711a351d5d9e9f068d23cc30147ca7a47"],"state_sha256":"ed16b06f5c186e77e248bff7cce35ab84c732aa27d1deb8dae6c7987d7e0d3e2"}