{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:FPF3GW2EQGY6NPZIXRYAOZPLMO","short_pith_number":"pith:FPF3GW2E","schema_version":"1.0","canonical_sha256":"2bcbb35b4481b1e6bf28bc700765eb6387f181c4fc737e1b4a724fa462292db1","source":{"kind":"arxiv","id":"2102.07301","version":2},"attestation_state":"computed","paper":{"title":"Nearly Minimax Optimal Regret for Learning Infinite-horizon Average-reward MDPs with Linear Function Approximation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["math.OC","stat.ML"],"primary_cat":"cs.LG","authors_text":"Dongruo Zhou, Quanquan Gu, Yue Wu","submitted_at":"2021-02-15T02:08:39Z","abstract_excerpt":"We study reinforcement learning in an infinite-horizon average-reward setting with linear function approximation, where the transition probability function of the underlying Markov Decision Process (MDP) admits a linear form over a feature mapping of the current state, action, and next state. We propose a new algorithm UCRL2-VTR, which can be seen as an extension of the UCRL2 algorithm with linear function approximation. We show that UCRL2-VTR with Bernstein-type bonus can achieve a regret of $\\tilde{O}(d\\sqrt{DT})$, where $d$ is the dimension of the feature mapping, $T$ is the horizon, and $\\"},"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":"2102.07301","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2021-02-15T02:08:39Z","cross_cats_sorted":["math.OC","stat.ML"],"title_canon_sha256":"04106b600bacdbd9a76f7a532bb76b8ecb056eb35ed11e706b6e1756e2ffc506","abstract_canon_sha256":"471054bc9a81e8f7576f37c5e463fa5b197859e55ac8420696cf42ea91736894"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T04:21:44.574424Z","signature_b64":"8WJClSKjNHOYxKXb8cLgmtsH6d2nVMtJc8/gQFzOjfUX6Bj65lJSIjaKFtoz/lDuJLVLZb6XHnNVCODWBm/UAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"2bcbb35b4481b1e6bf28bc700765eb6387f181c4fc737e1b4a724fa462292db1","last_reissued_at":"2026-07-05T04:21:44.573907Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T04:21:44.573907Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Nearly Minimax Optimal Regret for Learning Infinite-horizon Average-reward MDPs with Linear Function Approximation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["math.OC","stat.ML"],"primary_cat":"cs.LG","authors_text":"Dongruo Zhou, Quanquan Gu, Yue Wu","submitted_at":"2021-02-15T02:08:39Z","abstract_excerpt":"We study reinforcement learning in an infinite-horizon average-reward setting with linear function approximation, where the transition probability function of the underlying Markov Decision Process (MDP) admits a linear form over a feature mapping of the current state, action, and next state. We propose a new algorithm UCRL2-VTR, which can be seen as an extension of the UCRL2 algorithm with linear function approximation. We show that UCRL2-VTR with Bernstein-type bonus can achieve a regret of $\\tilde{O}(d\\sqrt{DT})$, where $d$ is the dimension of the feature mapping, $T$ is the horizon, and $\\"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2102.07301","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/2102.07301/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":"2102.07301","created_at":"2026-07-05T04:21:44.573969+00:00"},{"alias_kind":"arxiv_version","alias_value":"2102.07301v2","created_at":"2026-07-05T04:21:44.573969+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2102.07301","created_at":"2026-07-05T04:21:44.573969+00:00"},{"alias_kind":"pith_short_12","alias_value":"FPF3GW2EQGY6","created_at":"2026-07-05T04:21:44.573969+00:00"},{"alias_kind":"pith_short_16","alias_value":"FPF3GW2EQGY6NPZI","created_at":"2026-07-05T04:21:44.573969+00:00"},{"alias_kind":"pith_short_8","alias_value":"FPF3GW2E","created_at":"2026-07-05T04:21:44.573969+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2608.06545","citing_title":"Robust Average-Reward Markov Decision Processes: Minimax-Optimal Learning via Plug-in Reductions","ref_index":66,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/FPF3GW2EQGY6NPZIXRYAOZPLMO","json":"https://pith.science/pith/FPF3GW2EQGY6NPZIXRYAOZPLMO.json","graph_json":"https://pith.science/api/pith-number/FPF3GW2EQGY6NPZIXRYAOZPLMO/graph.json","events_json":"https://pith.science/api/pith-number/FPF3GW2EQGY6NPZIXRYAOZPLMO/events.json","paper":"https://pith.science/paper/FPF3GW2E"},"agent_actions":{"view_html":"https://pith.science/pith/FPF3GW2EQGY6NPZIXRYAOZPLMO","download_json":"https://pith.science/pith/FPF3GW2EQGY6NPZIXRYAOZPLMO.json","view_paper":"https://pith.science/paper/FPF3GW2E","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2102.07301&json=true","fetch_graph":"https://pith.science/api/pith-number/FPF3GW2EQGY6NPZIXRYAOZPLMO/graph.json","fetch_events":"https://pith.science/api/pith-number/FPF3GW2EQGY6NPZIXRYAOZPLMO/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/FPF3GW2EQGY6NPZIXRYAOZPLMO/action/timestamp_anchor","attest_storage":"https://pith.science/pith/FPF3GW2EQGY6NPZIXRYAOZPLMO/action/storage_attestation","attest_author":"https://pith.science/pith/FPF3GW2EQGY6NPZIXRYAOZPLMO/action/author_attestation","sign_citation":"https://pith.science/pith/FPF3GW2EQGY6NPZIXRYAOZPLMO/action/citation_signature","submit_replication":"https://pith.science/pith/FPF3GW2EQGY6NPZIXRYAOZPLMO/action/replication_record"}},"created_at":"2026-07-05T04:21:44.573969+00:00","updated_at":"2026-07-05T04:21:44.573969+00:00"}