{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:GKJ3GOISPNAS6TC6PP3GZJ6KLH","short_pith_number":"pith:GKJ3GOIS","schema_version":"1.0","canonical_sha256":"3293b339127b412f4c5e7bf66ca7ca59f7c043e8b8993004ab5d8e276c90ecd4","source":{"kind":"arxiv","id":"2302.05372","version":3},"attestation_state":"computed","paper":{"title":"Towards Minimax Optimality of Model-based Robust Reinforcement Learning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Erwan Le Pennec, Matthieu Geist, Pierre Clavier","submitted_at":"2023-02-10T16:50:40Z","abstract_excerpt":"We study the sample complexity of obtaining an $\\epsilon$-optimal policy in \\emph{Robust} discounted Markov Decision Processes (RMDPs), given only access to a generative model of the nominal kernel. This problem is widely studied in the non-robust case, and it is known that any planning approach applied to an empirical MDP estimated with $\\tilde{\\mathcal{O}}(\\frac{H^3 \\mid S \\mid\\mid A \\mid}{\\epsilon^2})$ samples provides an $\\epsilon$-optimal policy, which is minimax optimal. Results in the robust case are much more scarce. For $sa$- (resp $s$-)rectangular uncertainty sets, the best known sam"},"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":"2302.05372","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2023-02-10T16:50:40Z","cross_cats_sorted":["stat.ML"],"title_canon_sha256":"b780be6893245b98daecdcab62ac0145ac70461a53619644415f17e6be98b38a","abstract_canon_sha256":"5722e2184ad4425c3644f595f802748d9eb58ac0932d458f1ba918db49798763"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:28:00.186795Z","signature_b64":"DnjUmoa7l73rOElXqFpn24g7pvJ3BBGynjMeLvdGsbA2tmQkRGO6lbiHH8pHp7wRHUaT1MyjwyA8SOL5Omu0DQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"3293b339127b412f4c5e7bf66ca7ca59f7c043e8b8993004ab5d8e276c90ecd4","last_reissued_at":"2026-07-05T08:28:00.186398Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:28:00.186398Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Towards Minimax Optimality of Model-based Robust Reinforcement Learning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Erwan Le Pennec, Matthieu Geist, Pierre Clavier","submitted_at":"2023-02-10T16:50:40Z","abstract_excerpt":"We study the sample complexity of obtaining an $\\epsilon$-optimal policy in \\emph{Robust} discounted Markov Decision Processes (RMDPs), given only access to a generative model of the nominal kernel. This problem is widely studied in the non-robust case, and it is known that any planning approach applied to an empirical MDP estimated with $\\tilde{\\mathcal{O}}(\\frac{H^3 \\mid S \\mid\\mid A \\mid}{\\epsilon^2})$ samples provides an $\\epsilon$-optimal policy, which is minimax optimal. Results in the robust case are much more scarce. For $sa$- (resp $s$-)rectangular uncertainty sets, the best known sam"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2302.05372","kind":"arxiv","version":3},"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/2302.05372/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":"2302.05372","created_at":"2026-07-05T08:28:00.186452+00:00"},{"alias_kind":"arxiv_version","alias_value":"2302.05372v3","created_at":"2026-07-05T08:28:00.186452+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2302.05372","created_at":"2026-07-05T08:28:00.186452+00:00"},{"alias_kind":"pith_short_12","alias_value":"GKJ3GOISPNAS","created_at":"2026-07-05T08:28:00.186452+00:00"},{"alias_kind":"pith_short_16","alias_value":"GKJ3GOISPNAS6TC6","created_at":"2026-07-05T08:28:00.186452+00:00"},{"alias_kind":"pith_short_8","alias_value":"GKJ3GOIS","created_at":"2026-07-05T08:28:00.186452+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2408.16286","citing_title":"Near-Optimal Policy Identification in Robust Constrained Markov Decision Processes via Epigraph Form","ref_index":21,"is_internal_anchor":false},{"citing_arxiv_id":"2502.02844","citing_title":"Wolfpack Adversarial Attack for Robust Multi-Agent Reinforcement Learning","ref_index":1,"is_internal_anchor":false},{"citing_arxiv_id":"2506.12622","citing_title":"DR-SAC: Distributionally Robust Soft Actor-Critic for Reinforcement Learning under Uncertainty","ref_index":6,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/GKJ3GOISPNAS6TC6PP3GZJ6KLH","json":"https://pith.science/pith/GKJ3GOISPNAS6TC6PP3GZJ6KLH.json","graph_json":"https://pith.science/api/pith-number/GKJ3GOISPNAS6TC6PP3GZJ6KLH/graph.json","events_json":"https://pith.science/api/pith-number/GKJ3GOISPNAS6TC6PP3GZJ6KLH/events.json","paper":"https://pith.science/paper/GKJ3GOIS"},"agent_actions":{"view_html":"https://pith.science/pith/GKJ3GOISPNAS6TC6PP3GZJ6KLH","download_json":"https://pith.science/pith/GKJ3GOISPNAS6TC6PP3GZJ6KLH.json","view_paper":"https://pith.science/paper/GKJ3GOIS","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2302.05372&json=true","fetch_graph":"https://pith.science/api/pith-number/GKJ3GOISPNAS6TC6PP3GZJ6KLH/graph.json","fetch_events":"https://pith.science/api/pith-number/GKJ3GOISPNAS6TC6PP3GZJ6KLH/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/GKJ3GOISPNAS6TC6PP3GZJ6KLH/action/timestamp_anchor","attest_storage":"https://pith.science/pith/GKJ3GOISPNAS6TC6PP3GZJ6KLH/action/storage_attestation","attest_author":"https://pith.science/pith/GKJ3GOISPNAS6TC6PP3GZJ6KLH/action/author_attestation","sign_citation":"https://pith.science/pith/GKJ3GOISPNAS6TC6PP3GZJ6KLH/action/citation_signature","submit_replication":"https://pith.science/pith/GKJ3GOISPNAS6TC6PP3GZJ6KLH/action/replication_record"}},"created_at":"2026-07-05T08:28:00.186452+00:00","updated_at":"2026-07-05T08:28:00.186452+00:00"}