{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:IRPOU4KNJWMJHEA3EHYDGVRMMQ","short_pith_number":"pith:IRPOU4KN","schema_version":"1.0","canonical_sha256":"445eea714d4d9893901b21f033562c641c9a82627b2d683e3c1f45125a2ad420","source":{"kind":"arxiv","id":"2410.07574","version":2},"attestation_state":"computed","paper":{"title":"Gap-Dependent Bounds for Q-Learning using Reference-Advantage Decomposition","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":"2024-10-10T03:19:46Z","abstract_excerpt":"We study the gap-dependent bounds of two important algorithms for on-policy Q-learning for finite-horizon episodic tabular Markov Decision Processes (MDPs): UCB-Advantage (Zhang et al. 2020) and Q-EarlySettled-Advantage (Li et al. 2021). UCB-Advantage and Q-EarlySettled-Advantage improve upon the results based on Hoeffding-type bonuses and achieve the almost optimal $\\sqrt{T}$-type regret bound in the worst-case scenario, where $T$ is the total number of steps. However, the benign structures of the MDPs such as a strictly positive suboptimality gap can significantly improve the regret. While g"},"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":"2410.07574","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"stat.ML","submitted_at":"2024-10-10T03:19:46Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"3bb2b118e2d8b34c0b76fe31707991436c1e12ece974f55aa1ca3222f613a1db","abstract_canon_sha256":"e9c8587cd9a32058278ebec9ec17b21571ad282abbc035470459ca47d9c416c8"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:27:17.090545Z","signature_b64":"rNcOlHO5fO6BCavdt89VoMbgpwxnytrAChPzgvvnqTYnV1N5evwE1bWp6cecZppc/5cDB1rVQIjR/1JqD1buDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"445eea714d4d9893901b21f033562c641c9a82627b2d683e3c1f45125a2ad420","last_reissued_at":"2026-07-05T10:27:17.089518Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:27:17.089518Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Gap-Dependent Bounds for Q-Learning using Reference-Advantage Decomposition","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":"2024-10-10T03:19:46Z","abstract_excerpt":"We study the gap-dependent bounds of two important algorithms for on-policy Q-learning for finite-horizon episodic tabular Markov Decision Processes (MDPs): UCB-Advantage (Zhang et al. 2020) and Q-EarlySettled-Advantage (Li et al. 2021). UCB-Advantage and Q-EarlySettled-Advantage improve upon the results based on Hoeffding-type bonuses and achieve the almost optimal $\\sqrt{T}$-type regret bound in the worst-case scenario, where $T$ is the total number of steps. However, the benign structures of the MDPs such as a strictly positive suboptimality gap can significantly improve the regret. While g"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.07574","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/2410.07574/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":"2410.07574","created_at":"2026-07-05T10:27:17.089665+00:00"},{"alias_kind":"arxiv_version","alias_value":"2410.07574v2","created_at":"2026-07-05T10:27:17.089665+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.07574","created_at":"2026-07-05T10:27:17.089665+00:00"},{"alias_kind":"pith_short_12","alias_value":"IRPOU4KNJWMJ","created_at":"2026-07-05T10:27:17.089665+00:00"},{"alias_kind":"pith_short_16","alias_value":"IRPOU4KNJWMJHEA3","created_at":"2026-07-05T10:27:17.089665+00:00"},{"alias_kind":"pith_short_8","alias_value":"IRPOU4KN","created_at":"2026-07-05T10:27:17.089665+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/IRPOU4KNJWMJHEA3EHYDGVRMMQ","json":"https://pith.science/pith/IRPOU4KNJWMJHEA3EHYDGVRMMQ.json","graph_json":"https://pith.science/api/pith-number/IRPOU4KNJWMJHEA3EHYDGVRMMQ/graph.json","events_json":"https://pith.science/api/pith-number/IRPOU4KNJWMJHEA3EHYDGVRMMQ/events.json","paper":"https://pith.science/paper/IRPOU4KN"},"agent_actions":{"view_html":"https://pith.science/pith/IRPOU4KNJWMJHEA3EHYDGVRMMQ","download_json":"https://pith.science/pith/IRPOU4KNJWMJHEA3EHYDGVRMMQ.json","view_paper":"https://pith.science/paper/IRPOU4KN","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2410.07574&json=true","fetch_graph":"https://pith.science/api/pith-number/IRPOU4KNJWMJHEA3EHYDGVRMMQ/graph.json","fetch_events":"https://pith.science/api/pith-number/IRPOU4KNJWMJHEA3EHYDGVRMMQ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/IRPOU4KNJWMJHEA3EHYDGVRMMQ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/IRPOU4KNJWMJHEA3EHYDGVRMMQ/action/storage_attestation","attest_author":"https://pith.science/pith/IRPOU4KNJWMJHEA3EHYDGVRMMQ/action/author_attestation","sign_citation":"https://pith.science/pith/IRPOU4KNJWMJHEA3EHYDGVRMMQ/action/citation_signature","submit_replication":"https://pith.science/pith/IRPOU4KNJWMJHEA3EHYDGVRMMQ/action/replication_record"}},"created_at":"2026-07-05T10:27:17.089665+00:00","updated_at":"2026-07-05T10:27:17.089665+00:00"}