{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:DIOR6XV4RKBHDTKFB4T6CAQHYT","short_pith_number":"pith:DIOR6XV4","schema_version":"1.0","canonical_sha256":"1a1d1f5ebc8a8271cd450f27e10207c4df27fbc3a8e0b777ba9da3f916121b17","source":{"kind":"arxiv","id":"2504.05405","version":1},"attestation_state":"computed","paper":{"title":"The Role of Environment Access in Agnostic Reinforcement Learning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","stat.ML"],"primary_cat":"cs.LG","authors_text":"Akshay Krishnamurthy, Ayush Sekhari, Gene Li","submitted_at":"2025-04-07T18:19:56Z","abstract_excerpt":"We study Reinforcement Learning (RL) in environments with large state spaces, where function approximation is required for sample-efficient learning. Departing from a long history of prior work, we consider the weakest possible form of function approximation, called agnostic policy learning, where the learner seeks to find the best policy in a given class $\\Pi$, with no guarantee that $\\Pi$ contains an optimal policy for the underlying task. Although it is known that sample-efficient agnostic policy learning is not possible in the standard online RL setting without further assumptions, we inve"},"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":"2504.05405","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-04-07T18:19:56Z","cross_cats_sorted":["cs.AI","stat.ML"],"title_canon_sha256":"caad5bb620c2cd4635d57f4a05769e5857cc427db73a6b7d0b51ecbc3673cd42","abstract_canon_sha256":"be88433ef5c997eb1231ab0c051d6f7bb49e5181e5b5d1e2e3db524b5434026b"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:45:41.718441Z","signature_b64":"eVcV7zP2LXXezLYfCJTJNqDRddEZMpAYtpB3t67jDtI0YDAxy27/aQ9VX+YFvvcTVovlf2ckMhaNbLX9e6d9Aw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1a1d1f5ebc8a8271cd450f27e10207c4df27fbc3a8e0b777ba9da3f916121b17","last_reissued_at":"2026-07-05T10:45:41.717910Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:45:41.717910Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"The Role of Environment Access in Agnostic Reinforcement Learning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","stat.ML"],"primary_cat":"cs.LG","authors_text":"Akshay Krishnamurthy, Ayush Sekhari, Gene Li","submitted_at":"2025-04-07T18:19:56Z","abstract_excerpt":"We study Reinforcement Learning (RL) in environments with large state spaces, where function approximation is required for sample-efficient learning. Departing from a long history of prior work, we consider the weakest possible form of function approximation, called agnostic policy learning, where the learner seeks to find the best policy in a given class $\\Pi$, with no guarantee that $\\Pi$ contains an optimal policy for the underlying task. Although it is known that sample-efficient agnostic policy learning is not possible in the standard online RL setting without further assumptions, we inve"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2504.05405","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/2504.05405/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":"2504.05405","created_at":"2026-07-05T10:45:41.717978+00:00"},{"alias_kind":"arxiv_version","alias_value":"2504.05405v1","created_at":"2026-07-05T10:45:41.717978+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2504.05405","created_at":"2026-07-05T10:45:41.717978+00:00"},{"alias_kind":"pith_short_12","alias_value":"DIOR6XV4RKBH","created_at":"2026-07-05T10:45:41.717978+00:00"},{"alias_kind":"pith_short_16","alias_value":"DIOR6XV4RKBHDTKF","created_at":"2026-07-05T10:45:41.717978+00:00"},{"alias_kind":"pith_short_8","alias_value":"DIOR6XV4","created_at":"2026-07-05T10:45:41.717978+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2507.04406","citing_title":"Convergence and Sample Complexity of First-Order Methods for Agnostic Reinforcement Learning","ref_index":30,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/DIOR6XV4RKBHDTKFB4T6CAQHYT","json":"https://pith.science/pith/DIOR6XV4RKBHDTKFB4T6CAQHYT.json","graph_json":"https://pith.science/api/pith-number/DIOR6XV4RKBHDTKFB4T6CAQHYT/graph.json","events_json":"https://pith.science/api/pith-number/DIOR6XV4RKBHDTKFB4T6CAQHYT/events.json","paper":"https://pith.science/paper/DIOR6XV4"},"agent_actions":{"view_html":"https://pith.science/pith/DIOR6XV4RKBHDTKFB4T6CAQHYT","download_json":"https://pith.science/pith/DIOR6XV4RKBHDTKFB4T6CAQHYT.json","view_paper":"https://pith.science/paper/DIOR6XV4","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2504.05405&json=true","fetch_graph":"https://pith.science/api/pith-number/DIOR6XV4RKBHDTKFB4T6CAQHYT/graph.json","fetch_events":"https://pith.science/api/pith-number/DIOR6XV4RKBHDTKFB4T6CAQHYT/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/DIOR6XV4RKBHDTKFB4T6CAQHYT/action/timestamp_anchor","attest_storage":"https://pith.science/pith/DIOR6XV4RKBHDTKFB4T6CAQHYT/action/storage_attestation","attest_author":"https://pith.science/pith/DIOR6XV4RKBHDTKFB4T6CAQHYT/action/author_attestation","sign_citation":"https://pith.science/pith/DIOR6XV4RKBHDTKFB4T6CAQHYT/action/citation_signature","submit_replication":"https://pith.science/pith/DIOR6XV4RKBHDTKFB4T6CAQHYT/action/replication_record"}},"created_at":"2026-07-05T10:45:41.717978+00:00","updated_at":"2026-07-05T10:45:41.717978+00:00"}