{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:AOG3T77FGVVQCMRFHOCO3GQ6N5","short_pith_number":"pith:AOG3T77F","schema_version":"1.0","canonical_sha256":"038db9ffe5356b0132253b84ed9a1e6f41819b8b25211161f9ca876247e5b65b","source":{"kind":"arxiv","id":"2305.18443","version":1},"attestation_state":"computed","paper":{"title":"Off-Policy RL Algorithms Can be Sample-Efficient for Continuous Control via Sample Multiple Reuse","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Jiafei Lyu, Le Wan, Xiu Li, Zongqing Lu","submitted_at":"2023-05-29T03:25:22Z","abstract_excerpt":"Sample efficiency is one of the most critical issues for online reinforcement learning (RL). Existing methods achieve higher sample efficiency by adopting model-based methods, Q-ensemble, or better exploration mechanisms. We, instead, propose to train an off-policy RL agent via updating on a fixed sampled batch multiple times, thus reusing these samples and better exploiting them within a single optimization loop. We name our method sample multiple reuse (SMR). We theoretically show the properties of Q-learning with SMR, e.g., convergence. Furthermore, we incorporate SMR with off-the-shelf off"},"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":"2305.18443","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2023-05-29T03:25:22Z","cross_cats_sorted":[],"title_canon_sha256":"7b0c80249f63b7f85d5a71d391674b2c0776953d2a0f300915b31bc834f8e996","abstract_canon_sha256":"dd887a1737749b08f7967c6b98f089b16009d8abf5cc7d1d377c4826d89c4eb7"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T06:15:00.981307Z","signature_b64":"5SQFrGWX8PsYmA+G4gfFwHc5t9sA58KII2dgSOUWQ3y+AgvJaSIZZZcvNbZGFx6HeFlg72t2XZMCSmOyNB3TBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"038db9ffe5356b0132253b84ed9a1e6f41819b8b25211161f9ca876247e5b65b","last_reissued_at":"2026-07-05T06:15:00.980884Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T06:15:00.980884Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Off-Policy RL Algorithms Can be Sample-Efficient for Continuous Control via Sample Multiple Reuse","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Jiafei Lyu, Le Wan, Xiu Li, Zongqing Lu","submitted_at":"2023-05-29T03:25:22Z","abstract_excerpt":"Sample efficiency is one of the most critical issues for online reinforcement learning (RL). Existing methods achieve higher sample efficiency by adopting model-based methods, Q-ensemble, or better exploration mechanisms. We, instead, propose to train an off-policy RL agent via updating on a fixed sampled batch multiple times, thus reusing these samples and better exploiting them within a single optimization loop. We name our method sample multiple reuse (SMR). We theoretically show the properties of Q-learning with SMR, e.g., convergence. Furthermore, we incorporate SMR with off-the-shelf off"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2305.18443","kind":"arxiv","version":1},"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/2305.18443/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":"2305.18443","created_at":"2026-07-05T06:15:00.980940+00:00"},{"alias_kind":"arxiv_version","alias_value":"2305.18443v1","created_at":"2026-07-05T06:15:00.980940+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2305.18443","created_at":"2026-07-05T06:15:00.980940+00:00"},{"alias_kind":"pith_short_12","alias_value":"AOG3T77FGVVQ","created_at":"2026-07-05T06:15:00.980940+00:00"},{"alias_kind":"pith_short_16","alias_value":"AOG3T77FGVVQCMRF","created_at":"2026-07-05T06:15:00.980940+00:00"},{"alias_kind":"pith_short_8","alias_value":"AOG3T77F","created_at":"2026-07-05T06:15:00.980940+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2507.02712","citing_title":"A Forget-and-Grow Strategy for Deep Reinforcement Learning Scaling in Continuous Control","ref_index":27,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/AOG3T77FGVVQCMRFHOCO3GQ6N5","json":"https://pith.science/pith/AOG3T77FGVVQCMRFHOCO3GQ6N5.json","graph_json":"https://pith.science/api/pith-number/AOG3T77FGVVQCMRFHOCO3GQ6N5/graph.json","events_json":"https://pith.science/api/pith-number/AOG3T77FGVVQCMRFHOCO3GQ6N5/events.json","paper":"https://pith.science/paper/AOG3T77F"},"agent_actions":{"view_html":"https://pith.science/pith/AOG3T77FGVVQCMRFHOCO3GQ6N5","download_json":"https://pith.science/pith/AOG3T77FGVVQCMRFHOCO3GQ6N5.json","view_paper":"https://pith.science/paper/AOG3T77F","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2305.18443&json=true","fetch_graph":"https://pith.science/api/pith-number/AOG3T77FGVVQCMRFHOCO3GQ6N5/graph.json","fetch_events":"https://pith.science/api/pith-number/AOG3T77FGVVQCMRFHOCO3GQ6N5/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/AOG3T77FGVVQCMRFHOCO3GQ6N5/action/timestamp_anchor","attest_storage":"https://pith.science/pith/AOG3T77FGVVQCMRFHOCO3GQ6N5/action/storage_attestation","attest_author":"https://pith.science/pith/AOG3T77FGVVQCMRFHOCO3GQ6N5/action/author_attestation","sign_citation":"https://pith.science/pith/AOG3T77FGVVQCMRFHOCO3GQ6N5/action/citation_signature","submit_replication":"https://pith.science/pith/AOG3T77FGVVQCMRFHOCO3GQ6N5/action/replication_record"}},"created_at":"2026-07-05T06:15:00.980940+00:00","updated_at":"2026-07-05T06:15:00.980940+00:00"}