{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:76QB5FAYVB7HLHHOKYPSM7QY3G","short_pith_number":"pith:76QB5FAY","schema_version":"1.0","canonical_sha256":"ffa01e9418a87e759cee561f267e18d99366d2d8f22fd9a0a85529b1b76f4899","source":{"kind":"arxiv","id":"2105.08140","version":1},"attestation_state":"computed","paper":{"title":"Uncertainty Weighted Actor-Critic for Offline Reinforcement Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Hanlin Goh, Jian Zhang, Joshua Susskind, Nitish Srivastava, Ruslan Salakhutdinov, Shuangfei Zhai, Yue Wu","submitted_at":"2021-05-17T20:16:46Z","abstract_excerpt":"Offline Reinforcement Learning promises to learn effective policies from previously-collected, static datasets without the need for exploration. However, existing Q-learning and actor-critic based off-policy RL algorithms fail when bootstrapping from out-of-distribution (OOD) actions or states. We hypothesize that a key missing ingredient from the existing methods is a proper treatment of uncertainty in the offline setting. We propose Uncertainty Weighted Actor-Critic (UWAC), an algorithm that detects OOD state-action pairs and down-weights their contribution in the training objectives accordi"},"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":"2105.08140","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2021-05-17T20:16:46Z","cross_cats_sorted":[],"title_canon_sha256":"2c6a1945d35e792167ab35201fcc2b6e1c7f022657afa5b76332eb16cf5ef0a2","abstract_canon_sha256":"84526f6b4838fd3a63e3d0ba0e0d056121e8c5074caff5b2d88e2fc578712e89"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T02:40:52.042258Z","signature_b64":"nyO9OKC0Jw73C9eFiZk7r6YAwzcBikwzdL7P+sKDS1RbfbthSUqzrzSd2eMSVDY7aw5tjv9R4ba9MOwO5VReAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"ffa01e9418a87e759cee561f267e18d99366d2d8f22fd9a0a85529b1b76f4899","last_reissued_at":"2026-07-05T02:40:52.041701Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T02:40:52.041701Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Uncertainty Weighted Actor-Critic for Offline Reinforcement Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Hanlin Goh, Jian Zhang, Joshua Susskind, Nitish Srivastava, Ruslan Salakhutdinov, Shuangfei Zhai, Yue Wu","submitted_at":"2021-05-17T20:16:46Z","abstract_excerpt":"Offline Reinforcement Learning promises to learn effective policies from previously-collected, static datasets without the need for exploration. However, existing Q-learning and actor-critic based off-policy RL algorithms fail when bootstrapping from out-of-distribution (OOD) actions or states. We hypothesize that a key missing ingredient from the existing methods is a proper treatment of uncertainty in the offline setting. We propose Uncertainty Weighted Actor-Critic (UWAC), an algorithm that detects OOD state-action pairs and down-weights their contribution in the training objectives accordi"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2105.08140","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/2105.08140/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":"2105.08140","created_at":"2026-07-05T02:40:52.041765+00:00"},{"alias_kind":"arxiv_version","alias_value":"2105.08140v1","created_at":"2026-07-05T02:40:52.041765+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2105.08140","created_at":"2026-07-05T02:40:52.041765+00:00"},{"alias_kind":"pith_short_12","alias_value":"76QB5FAYVB7H","created_at":"2026-07-05T02:40:52.041765+00:00"},{"alias_kind":"pith_short_16","alias_value":"76QB5FAYVB7HLHHO","created_at":"2026-07-05T02:40:52.041765+00:00"},{"alias_kind":"pith_short_8","alias_value":"76QB5FAY","created_at":"2026-07-05T02:40:52.041765+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.07592","citing_title":"UNIQ: Conformal Calibration for Adaptive Conservatism in Offline Reinforcement Learning","ref_index":30,"is_internal_anchor":false},{"citing_arxiv_id":"2605.08202","citing_title":"Beyond Penalization: Diffusion-based Out-of-Distribution Detection and Selective Regularization in Offline Reinforcement Learning","ref_index":10,"is_internal_anchor":false},{"citing_arxiv_id":"2605.01663","citing_title":"Towards Efficient and Expressive Offline RL via Flow-Anchored Noise-conditioned Q-Learning","ref_index":33,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/76QB5FAYVB7HLHHOKYPSM7QY3G","json":"https://pith.science/pith/76QB5FAYVB7HLHHOKYPSM7QY3G.json","graph_json":"https://pith.science/api/pith-number/76QB5FAYVB7HLHHOKYPSM7QY3G/graph.json","events_json":"https://pith.science/api/pith-number/76QB5FAYVB7HLHHOKYPSM7QY3G/events.json","paper":"https://pith.science/paper/76QB5FAY"},"agent_actions":{"view_html":"https://pith.science/pith/76QB5FAYVB7HLHHOKYPSM7QY3G","download_json":"https://pith.science/pith/76QB5FAYVB7HLHHOKYPSM7QY3G.json","view_paper":"https://pith.science/paper/76QB5FAY","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2105.08140&json=true","fetch_graph":"https://pith.science/api/pith-number/76QB5FAYVB7HLHHOKYPSM7QY3G/graph.json","fetch_events":"https://pith.science/api/pith-number/76QB5FAYVB7HLHHOKYPSM7QY3G/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/76QB5FAYVB7HLHHOKYPSM7QY3G/action/timestamp_anchor","attest_storage":"https://pith.science/pith/76QB5FAYVB7HLHHOKYPSM7QY3G/action/storage_attestation","attest_author":"https://pith.science/pith/76QB5FAYVB7HLHHOKYPSM7QY3G/action/author_attestation","sign_citation":"https://pith.science/pith/76QB5FAYVB7HLHHOKYPSM7QY3G/action/citation_signature","submit_replication":"https://pith.science/pith/76QB5FAYVB7HLHHOKYPSM7QY3G/action/replication_record"}},"created_at":"2026-07-05T02:40:52.041765+00:00","updated_at":"2026-07-05T02:40:52.041765+00:00"}