{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:D74UIN55W4FCUQM2Q5AOWYOBM5","short_pith_number":"pith:D74UIN55","schema_version":"1.0","canonical_sha256":"1ff94437bdb70a2a419a8740eb61c16742c9f8e0c0aba016e36c28dca6f78e26","source":{"kind":"arxiv","id":"2505.11478","version":2},"attestation_state":"computed","paper":{"title":"Automatic Reward Shaping from Confounded Offline Data","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.AI","authors_text":"Elias Bareinboim, Junzhe Zhang, Mingxuan Li","submitted_at":"2025-05-16T17:40:01Z","abstract_excerpt":"A key task in Artificial Intelligence is learning effective policies for controlling agents in unknown environments to optimize performance measures. Off-policy learning methods, like Q-learning, allow learners to make optimal decisions based on past experiences. This paper studies off-policy learning from biased data in complex and high-dimensional domains where \\emph{unobserved confounding} cannot be ruled out a priori. Building on the well-celebrated Deep Q-Network (DQN), we propose a novel deep reinforcement learning algorithm robust to confounding biases in observed data. Specifically, ou"},"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":"2505.11478","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2025-05-16T17:40:01Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"3d7fdb6aed13e19f27af651e79bb7a1bb89553f213d404c6d0fe5b649e2ef33f","abstract_canon_sha256":"70b53be951adb6677d0bec43d882455c0d5da62d4fd33e84916bbff87c440fa6"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T12:07:38.419107Z","signature_b64":"UbwsZGOTPjEr85YbQoa6c9zPTd/jBhVl1y9ROdq/tFGVmBJfSrzDXCVuiJCfUctrVBfEIDytOhIBgZdFF/77CQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1ff94437bdb70a2a419a8740eb61c16742c9f8e0c0aba016e36c28dca6f78e26","last_reissued_at":"2026-07-05T12:07:38.418417Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T12:07:38.418417Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Automatic Reward Shaping from Confounded Offline Data","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.AI","authors_text":"Elias Bareinboim, Junzhe Zhang, Mingxuan Li","submitted_at":"2025-05-16T17:40:01Z","abstract_excerpt":"A key task in Artificial Intelligence is learning effective policies for controlling agents in unknown environments to optimize performance measures. Off-policy learning methods, like Q-learning, allow learners to make optimal decisions based on past experiences. This paper studies off-policy learning from biased data in complex and high-dimensional domains where \\emph{unobserved confounding} cannot be ruled out a priori. Building on the well-celebrated Deep Q-Network (DQN), we propose a novel deep reinforcement learning algorithm robust to confounding biases in observed data. Specifically, ou"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.11478","kind":"arxiv","version":2},"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/2505.11478/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":"2505.11478","created_at":"2026-07-05T12:07:38.418523+00:00"},{"alias_kind":"arxiv_version","alias_value":"2505.11478v2","created_at":"2026-07-05T12:07:38.418523+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.11478","created_at":"2026-07-05T12:07:38.418523+00:00"},{"alias_kind":"pith_short_12","alias_value":"D74UIN55W4FC","created_at":"2026-07-05T12:07:38.418523+00:00"},{"alias_kind":"pith_short_16","alias_value":"D74UIN55W4FCUQM2","created_at":"2026-07-05T12:07:38.418523+00:00"},{"alias_kind":"pith_short_8","alias_value":"D74UIN55","created_at":"2026-07-05T12:07:38.418523+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.07057","citing_title":"Integrating Causal DAGs in Deep RL: Activating Minimal Markovian States with Multi-Order Exposure","ref_index":15,"is_internal_anchor":false},{"citing_arxiv_id":"2604.20627","citing_title":"Occupancy Reward Shaping: Improving Credit Assignment for Offline Goal-Conditioned Reinforcement Learning","ref_index":11,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/D74UIN55W4FCUQM2Q5AOWYOBM5","json":"https://pith.science/pith/D74UIN55W4FCUQM2Q5AOWYOBM5.json","graph_json":"https://pith.science/api/pith-number/D74UIN55W4FCUQM2Q5AOWYOBM5/graph.json","events_json":"https://pith.science/api/pith-number/D74UIN55W4FCUQM2Q5AOWYOBM5/events.json","paper":"https://pith.science/paper/D74UIN55"},"agent_actions":{"view_html":"https://pith.science/pith/D74UIN55W4FCUQM2Q5AOWYOBM5","download_json":"https://pith.science/pith/D74UIN55W4FCUQM2Q5AOWYOBM5.json","view_paper":"https://pith.science/paper/D74UIN55","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2505.11478&json=true","fetch_graph":"https://pith.science/api/pith-number/D74UIN55W4FCUQM2Q5AOWYOBM5/graph.json","fetch_events":"https://pith.science/api/pith-number/D74UIN55W4FCUQM2Q5AOWYOBM5/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/D74UIN55W4FCUQM2Q5AOWYOBM5/action/timestamp_anchor","attest_storage":"https://pith.science/pith/D74UIN55W4FCUQM2Q5AOWYOBM5/action/storage_attestation","attest_author":"https://pith.science/pith/D74UIN55W4FCUQM2Q5AOWYOBM5/action/author_attestation","sign_citation":"https://pith.science/pith/D74UIN55W4FCUQM2Q5AOWYOBM5/action/citation_signature","submit_replication":"https://pith.science/pith/D74UIN55W4FCUQM2Q5AOWYOBM5/action/replication_record"}},"created_at":"2026-07-05T12:07:38.418523+00:00","updated_at":"2026-07-05T12:07:38.418523+00:00"}