{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:NDNTE6MCXWTRADT44BHNZ465AN","short_pith_number":"pith:NDNTE6MC","schema_version":"1.0","canonical_sha256":"68db327982bda7100e7ce04edcf3dd03585ec7b897b19801e358cf3ecace6d74","source":{"kind":"arxiv","id":"2405.18795","version":2},"attestation_state":"computed","paper":{"title":"Federated Q-Learning with Reference-Advantage Decomposition: Almost Optimal Regret and Logarithmic Communication Cost","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"stat.ML","authors_text":"Haochen Zhang, Lingzhou Xue, Zhong Zheng","submitted_at":"2024-05-29T06:26:52Z","abstract_excerpt":"In this paper, we consider model-free federated reinforcement learning for tabular episodic Markov decision processes. Under the coordination of a central server, multiple agents collaboratively explore the environment and learn an optimal policy without sharing their raw data. Despite recent advances in federated Q-learning algorithms achieving near-linear regret speedup with low communication cost, existing algorithms only attain suboptimal regrets compared to the information bound. We propose a novel model-free federated Q-learning algorithm, termed FedQ-Advantage. Our algorithm leverages r"},"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":"2405.18795","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"stat.ML","submitted_at":"2024-05-29T06:26:52Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"6124a2ac898f922fb3beb41cb8ca72a3d462cc4181769a391633f2dab5213741","abstract_canon_sha256":"ff3d2d6785517dea98dd1545bded4ec9528e73bf2e765458cc9f6de7c8b12e73"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:27:10.657166Z","signature_b64":"bfTWZEFhUNW1Ph7Hd8VxOZawlImZxoygin7D7gJA2K6pcgJrQ1yyw37lyDy5QUKqXLQwwyMocSZVWfhAqgPtDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"68db327982bda7100e7ce04edcf3dd03585ec7b897b19801e358cf3ecace6d74","last_reissued_at":"2026-07-05T10:27:10.656677Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:27:10.656677Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Federated Q-Learning with Reference-Advantage Decomposition: Almost Optimal Regret and Logarithmic Communication Cost","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"stat.ML","authors_text":"Haochen Zhang, Lingzhou Xue, Zhong Zheng","submitted_at":"2024-05-29T06:26:52Z","abstract_excerpt":"In this paper, we consider model-free federated reinforcement learning for tabular episodic Markov decision processes. Under the coordination of a central server, multiple agents collaboratively explore the environment and learn an optimal policy without sharing their raw data. Despite recent advances in federated Q-learning algorithms achieving near-linear regret speedup with low communication cost, existing algorithms only attain suboptimal regrets compared to the information bound. We propose a novel model-free federated Q-learning algorithm, termed FedQ-Advantage. Our algorithm leverages r"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2405.18795","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/2405.18795/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":"2405.18795","created_at":"2026-07-05T10:27:10.656738+00:00"},{"alias_kind":"arxiv_version","alias_value":"2405.18795v2","created_at":"2026-07-05T10:27:10.656738+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2405.18795","created_at":"2026-07-05T10:27:10.656738+00:00"},{"alias_kind":"pith_short_12","alias_value":"NDNTE6MCXWTR","created_at":"2026-07-05T10:27:10.656738+00:00"},{"alias_kind":"pith_short_16","alias_value":"NDNTE6MCXWTRADT4","created_at":"2026-07-05T10:27:10.656738+00:00"},{"alias_kind":"pith_short_8","alias_value":"NDNTE6MC","created_at":"2026-07-05T10:27:10.656738+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.08378","citing_title":"Reinforcement Learning for Scalable and Trustworthy Intelligent Systems","ref_index":69,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/NDNTE6MCXWTRADT44BHNZ465AN","json":"https://pith.science/pith/NDNTE6MCXWTRADT44BHNZ465AN.json","graph_json":"https://pith.science/api/pith-number/NDNTE6MCXWTRADT44BHNZ465AN/graph.json","events_json":"https://pith.science/api/pith-number/NDNTE6MCXWTRADT44BHNZ465AN/events.json","paper":"https://pith.science/paper/NDNTE6MC"},"agent_actions":{"view_html":"https://pith.science/pith/NDNTE6MCXWTRADT44BHNZ465AN","download_json":"https://pith.science/pith/NDNTE6MCXWTRADT44BHNZ465AN.json","view_paper":"https://pith.science/paper/NDNTE6MC","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2405.18795&json=true","fetch_graph":"https://pith.science/api/pith-number/NDNTE6MCXWTRADT44BHNZ465AN/graph.json","fetch_events":"https://pith.science/api/pith-number/NDNTE6MCXWTRADT44BHNZ465AN/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/NDNTE6MCXWTRADT44BHNZ465AN/action/timestamp_anchor","attest_storage":"https://pith.science/pith/NDNTE6MCXWTRADT44BHNZ465AN/action/storage_attestation","attest_author":"https://pith.science/pith/NDNTE6MCXWTRADT44BHNZ465AN/action/author_attestation","sign_citation":"https://pith.science/pith/NDNTE6MCXWTRADT44BHNZ465AN/action/citation_signature","submit_replication":"https://pith.science/pith/NDNTE6MCXWTRADT44BHNZ465AN/action/replication_record"}},"created_at":"2026-07-05T10:27:10.656738+00:00","updated_at":"2026-07-05T10:27:10.656738+00:00"}