{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:BA5XDTTSAQBO4J5JROXASYKFKK","short_pith_number":"pith:BA5XDTTS","schema_version":"1.0","canonical_sha256":"083b71ce720402ee27a98bae0961455293b1ed1aa042e08ec16635b8a1ab3652","source":{"kind":"arxiv","id":"2303.05606","version":2},"attestation_state":"computed","paper":{"title":"Variance-aware robust reinforcement learning with linear function approximation under heavy-tailed rewards","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","math.ST","stat.ML","stat.TH"],"primary_cat":"cs.LG","authors_text":"Qiang Sun, Xiang Li","submitted_at":"2023-03-09T22:16:28Z","abstract_excerpt":"This paper presents two algorithms, AdaOFUL and VARA, for online sequential decision-making in the presence of heavy-tailed rewards with only finite variances. For linear stochastic bandits, we address the issue of heavy-tailed rewards by modifying the adaptive Huber regression and proposing AdaOFUL. AdaOFUL achieves a state-of-the-art regret bound of $\\widetilde{O}\\big(d\\big(\\sum_{t=1}^T \\nu_{t}^2\\big)^{1/2}+d\\big)$ as if the rewards were uniformly bounded, where $\\nu_{t}^2$ is the observed conditional variance of the reward at round $t$, $d$ is the feature dimension, and $\\widetilde{O}(\\cdot"},"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":"2303.05606","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2023-03-09T22:16:28Z","cross_cats_sorted":["cs.AI","math.ST","stat.ML","stat.TH"],"title_canon_sha256":"535ed8d63f0ce946b79b51a35386f5697ea30c6dce88c73b050dc5ddb1737201","abstract_canon_sha256":"a916f7c573b704c448855f210ae69a0301b2fafb83f8768d10ed7e0c3f6611ca"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T05:51:04.328226Z","signature_b64":"q+Q4aqyPzEmyP8Okha7kG1ydSNKlOVb2yIOu5/zc0cW2iQenCRLKzyTF7WDshHZT24cdQ0af63F04Phze7LCDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"083b71ce720402ee27a98bae0961455293b1ed1aa042e08ec16635b8a1ab3652","last_reissued_at":"2026-07-05T05:51:04.327768Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T05:51:04.327768Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Variance-aware robust reinforcement learning with linear function approximation under heavy-tailed rewards","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","math.ST","stat.ML","stat.TH"],"primary_cat":"cs.LG","authors_text":"Qiang Sun, Xiang Li","submitted_at":"2023-03-09T22:16:28Z","abstract_excerpt":"This paper presents two algorithms, AdaOFUL and VARA, for online sequential decision-making in the presence of heavy-tailed rewards with only finite variances. For linear stochastic bandits, we address the issue of heavy-tailed rewards by modifying the adaptive Huber regression and proposing AdaOFUL. AdaOFUL achieves a state-of-the-art regret bound of $\\widetilde{O}\\big(d\\big(\\sum_{t=1}^T \\nu_{t}^2\\big)^{1/2}+d\\big)$ as if the rewards were uniformly bounded, where $\\nu_{t}^2$ is the observed conditional variance of the reward at round $t$, $d$ is the feature dimension, and $\\widetilde{O}(\\cdot"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2303.05606","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/2303.05606/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":"2303.05606","created_at":"2026-07-05T05:51:04.327825+00:00"},{"alias_kind":"arxiv_version","alias_value":"2303.05606v2","created_at":"2026-07-05T05:51:04.327825+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2303.05606","created_at":"2026-07-05T05:51:04.327825+00:00"},{"alias_kind":"pith_short_12","alias_value":"BA5XDTTSAQBO","created_at":"2026-07-05T05:51:04.327825+00:00"},{"alias_kind":"pith_short_16","alias_value":"BA5XDTTSAQBO4J5J","created_at":"2026-07-05T05:51:04.327825+00:00"},{"alias_kind":"pith_short_8","alias_value":"BA5XDTTS","created_at":"2026-07-05T05:51:04.327825+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2505.12037","citing_title":"Adaptive Resolving Methods for Reinforcement Learning with Function Approximations","ref_index":36,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/BA5XDTTSAQBO4J5JROXASYKFKK","json":"https://pith.science/pith/BA5XDTTSAQBO4J5JROXASYKFKK.json","graph_json":"https://pith.science/api/pith-number/BA5XDTTSAQBO4J5JROXASYKFKK/graph.json","events_json":"https://pith.science/api/pith-number/BA5XDTTSAQBO4J5JROXASYKFKK/events.json","paper":"https://pith.science/paper/BA5XDTTS"},"agent_actions":{"view_html":"https://pith.science/pith/BA5XDTTSAQBO4J5JROXASYKFKK","download_json":"https://pith.science/pith/BA5XDTTSAQBO4J5JROXASYKFKK.json","view_paper":"https://pith.science/paper/BA5XDTTS","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2303.05606&json=true","fetch_graph":"https://pith.science/api/pith-number/BA5XDTTSAQBO4J5JROXASYKFKK/graph.json","fetch_events":"https://pith.science/api/pith-number/BA5XDTTSAQBO4J5JROXASYKFKK/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/BA5XDTTSAQBO4J5JROXASYKFKK/action/timestamp_anchor","attest_storage":"https://pith.science/pith/BA5XDTTSAQBO4J5JROXASYKFKK/action/storage_attestation","attest_author":"https://pith.science/pith/BA5XDTTSAQBO4J5JROXASYKFKK/action/author_attestation","sign_citation":"https://pith.science/pith/BA5XDTTSAQBO4J5JROXASYKFKK/action/citation_signature","submit_replication":"https://pith.science/pith/BA5XDTTSAQBO4J5JROXASYKFKK/action/replication_record"}},"created_at":"2026-07-05T05:51:04.327825+00:00","updated_at":"2026-07-05T05:51:04.327825+00:00"}