{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2020:YAO2B6E6IO2HQIVTS4L7GDZ4OZ","short_pith_number":"pith:YAO2B6E6","schema_version":"1.0","canonical_sha256":"c01da0f89e43b47822b39717f30f3c7678f44dc30b5c29809b0cbc1821f27220","source":{"kind":"arxiv","id":"2011.04709","version":2},"attestation_state":"computed","paper":{"title":"f-IRL: Inverse Reinforcement Learning via State Marginal Matching","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.RO"],"primary_cat":"cs.LG","authors_text":"Benjamin Eysenbach, Harshit Sikchi, Lisa Lee, Tejus Gupta, Tianwei Ni, Yufei Wang","submitted_at":"2020-11-09T19:37:48Z","abstract_excerpt":"Imitation learning is well-suited for robotic tasks where it is difficult to directly program the behavior or specify a cost for optimal control. In this work, we propose a method for learning the reward function (and the corresponding policy) to match the expert state density. Our main result is the analytic gradient of any f-divergence between the agent and expert state distribution w.r.t. reward parameters. Based on the derived gradient, we present an algorithm, f-IRL, that recovers a stationary reward function from the expert density by gradient descent. We show that f-IRL can learn behavi"},"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":"2011.04709","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2020-11-09T19:37:48Z","cross_cats_sorted":["cs.RO"],"title_canon_sha256":"bf14a4a95251c2b81fe322d2957e4dd6f6cf1ea8ad375338a18597eaa6ad804e","abstract_canon_sha256":"a81b61f184118e5d32377e6f25f505c0074acd532d0a7dad3e7ed71dec3fc07d"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T02:02:56.921448Z","signature_b64":"IPcUigiE7rmV8WgZYHtNtwyySYKUXIvCc8r54p+6LQ9XAvW72b0rg/IGDsf816wfnmciTPZpbi4Y31rZHlZ9Bw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c01da0f89e43b47822b39717f30f3c7678f44dc30b5c29809b0cbc1821f27220","last_reissued_at":"2026-07-05T02:02:56.921064Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T02:02:56.921064Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"f-IRL: Inverse Reinforcement Learning via State Marginal Matching","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.RO"],"primary_cat":"cs.LG","authors_text":"Benjamin Eysenbach, Harshit Sikchi, Lisa Lee, Tejus Gupta, Tianwei Ni, Yufei Wang","submitted_at":"2020-11-09T19:37:48Z","abstract_excerpt":"Imitation learning is well-suited for robotic tasks where it is difficult to directly program the behavior or specify a cost for optimal control. In this work, we propose a method for learning the reward function (and the corresponding policy) to match the expert state density. Our main result is the analytic gradient of any f-divergence between the agent and expert state distribution w.r.t. reward parameters. Based on the derived gradient, we present an algorithm, f-IRL, that recovers a stationary reward function from the expert density by gradient descent. We show that f-IRL can learn behavi"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2011.04709","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/2011.04709/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":"2011.04709","created_at":"2026-07-05T02:02:56.921130+00:00"},{"alias_kind":"arxiv_version","alias_value":"2011.04709v2","created_at":"2026-07-05T02:02:56.921130+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2011.04709","created_at":"2026-07-05T02:02:56.921130+00:00"},{"alias_kind":"pith_short_12","alias_value":"YAO2B6E6IO2H","created_at":"2026-07-05T02:02:56.921130+00:00"},{"alias_kind":"pith_short_16","alias_value":"YAO2B6E6IO2HQIVT","created_at":"2026-07-05T02:02:56.921130+00:00"},{"alias_kind":"pith_short_8","alias_value":"YAO2B6E6","created_at":"2026-07-05T02:02:56.921130+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/YAO2B6E6IO2HQIVTS4L7GDZ4OZ","json":"https://pith.science/pith/YAO2B6E6IO2HQIVTS4L7GDZ4OZ.json","graph_json":"https://pith.science/api/pith-number/YAO2B6E6IO2HQIVTS4L7GDZ4OZ/graph.json","events_json":"https://pith.science/api/pith-number/YAO2B6E6IO2HQIVTS4L7GDZ4OZ/events.json","paper":"https://pith.science/paper/YAO2B6E6"},"agent_actions":{"view_html":"https://pith.science/pith/YAO2B6E6IO2HQIVTS4L7GDZ4OZ","download_json":"https://pith.science/pith/YAO2B6E6IO2HQIVTS4L7GDZ4OZ.json","view_paper":"https://pith.science/paper/YAO2B6E6","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2011.04709&json=true","fetch_graph":"https://pith.science/api/pith-number/YAO2B6E6IO2HQIVTS4L7GDZ4OZ/graph.json","fetch_events":"https://pith.science/api/pith-number/YAO2B6E6IO2HQIVTS4L7GDZ4OZ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/YAO2B6E6IO2HQIVTS4L7GDZ4OZ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/YAO2B6E6IO2HQIVTS4L7GDZ4OZ/action/storage_attestation","attest_author":"https://pith.science/pith/YAO2B6E6IO2HQIVTS4L7GDZ4OZ/action/author_attestation","sign_citation":"https://pith.science/pith/YAO2B6E6IO2HQIVTS4L7GDZ4OZ/action/citation_signature","submit_replication":"https://pith.science/pith/YAO2B6E6IO2HQIVTS4L7GDZ4OZ/action/replication_record"}},"created_at":"2026-07-05T02:02:56.921130+00:00","updated_at":"2026-07-05T02:02:56.921130+00:00"}