{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:7K6RHFINCJS74RMQXOUXYLXOJP","short_pith_number":"pith:7K6RHFIN","schema_version":"1.0","canonical_sha256":"fabd13950d1265fe4590bba97c2eee4bc595ec25a3984913c1c84bf9ae335063","source":{"kind":"arxiv","id":"2506.05445","version":1},"attestation_state":"computed","paper":{"title":"Causal Policy Learning in Reinforcement Learning: Backdoor-Adjusted Soft Actor-Critic","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Chien Lu, Haozhe Ma, Thanh Vinh Vo, Tze-Yun Leong, Young Lee","submitted_at":"2025-06-05T13:52:38Z","abstract_excerpt":"Hidden confounders that influence both states and actions can bias policy learning in reinforcement learning (RL), leading to suboptimal or non-generalizable behavior. Most RL algorithms ignore this issue, learning policies from observational trajectories based solely on statistical associations rather than causal effects. We propose DoSAC (Do-Calculus Soft Actor-Critic with Backdoor Adjustment), a principled extension of the SAC algorithm that corrects for hidden confounding via causal intervention estimation. DoSAC estimates the interventional policy $\\pi(a | \\mathrm{do}(s))$ using the backd"},"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":"2506.05445","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2025-06-05T13:52:38Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"c71b1f74224333cde8f4f81cc624abef639d8093bce8dadc1134854c30b259f5","abstract_canon_sha256":"490a06086f0bb93d155a76e969879807f2f4fd453ece3af00b65537d27b72d19"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:17:02.849107Z","signature_b64":"pcZ0jlB1FEth3rWJt+nXD3RXLA9glg+9FLzPVUseJYS9U4deTUdDm9W5IMx2Uf/MLI6iroopootcwCsQZ5dXAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"fabd13950d1265fe4590bba97c2eee4bc595ec25a3984913c1c84bf9ae335063","last_reissued_at":"2026-07-05T11:17:02.846790Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:17:02.846790Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Causal Policy Learning in Reinforcement Learning: Backdoor-Adjusted Soft Actor-Critic","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Chien Lu, Haozhe Ma, Thanh Vinh Vo, Tze-Yun Leong, Young Lee","submitted_at":"2025-06-05T13:52:38Z","abstract_excerpt":"Hidden confounders that influence both states and actions can bias policy learning in reinforcement learning (RL), leading to suboptimal or non-generalizable behavior. Most RL algorithms ignore this issue, learning policies from observational trajectories based solely on statistical associations rather than causal effects. We propose DoSAC (Do-Calculus Soft Actor-Critic with Backdoor Adjustment), a principled extension of the SAC algorithm that corrects for hidden confounding via causal intervention estimation. DoSAC estimates the interventional policy $\\pi(a | \\mathrm{do}(s))$ using the backd"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.05445","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/2506.05445/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":"2506.05445","created_at":"2026-07-05T11:17:02.848578+00:00"},{"alias_kind":"arxiv_version","alias_value":"2506.05445v1","created_at":"2026-07-05T11:17:02.848578+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.05445","created_at":"2026-07-05T11:17:02.848578+00:00"},{"alias_kind":"pith_short_12","alias_value":"7K6RHFINCJS7","created_at":"2026-07-05T11:17:02.848578+00:00"},{"alias_kind":"pith_short_16","alias_value":"7K6RHFINCJS74RMQ","created_at":"2026-07-05T11:17:02.848578+00:00"},{"alias_kind":"pith_short_8","alias_value":"7K6RHFIN","created_at":"2026-07-05T11:17:02.848578+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/7K6RHFINCJS74RMQXOUXYLXOJP","json":"https://pith.science/pith/7K6RHFINCJS74RMQXOUXYLXOJP.json","graph_json":"https://pith.science/api/pith-number/7K6RHFINCJS74RMQXOUXYLXOJP/graph.json","events_json":"https://pith.science/api/pith-number/7K6RHFINCJS74RMQXOUXYLXOJP/events.json","paper":"https://pith.science/paper/7K6RHFIN"},"agent_actions":{"view_html":"https://pith.science/pith/7K6RHFINCJS74RMQXOUXYLXOJP","download_json":"https://pith.science/pith/7K6RHFINCJS74RMQXOUXYLXOJP.json","view_paper":"https://pith.science/paper/7K6RHFIN","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2506.05445&json=true","fetch_graph":"https://pith.science/api/pith-number/7K6RHFINCJS74RMQXOUXYLXOJP/graph.json","fetch_events":"https://pith.science/api/pith-number/7K6RHFINCJS74RMQXOUXYLXOJP/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/7K6RHFINCJS74RMQXOUXYLXOJP/action/timestamp_anchor","attest_storage":"https://pith.science/pith/7K6RHFINCJS74RMQXOUXYLXOJP/action/storage_attestation","attest_author":"https://pith.science/pith/7K6RHFINCJS74RMQXOUXYLXOJP/action/author_attestation","sign_citation":"https://pith.science/pith/7K6RHFINCJS74RMQXOUXYLXOJP/action/citation_signature","submit_replication":"https://pith.science/pith/7K6RHFINCJS74RMQXOUXYLXOJP/action/replication_record"}},"created_at":"2026-07-05T11:17:02.848578+00:00","updated_at":"2026-07-05T11:17:02.848578+00:00"}