{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:EWFWDWNKVAIWMNIUC2QEUNTLIQ","short_pith_number":"pith:EWFWDWNK","schema_version":"1.0","canonical_sha256":"258b61d9aaa81166351416a04a366b4427e7d274d326c1af4c6c4ec89e595855","source":{"kind":"arxiv","id":"2308.02223","version":1},"attestation_state":"computed","paper":{"title":"ESRL: Efficient Sampling-based Reinforcement Learning for Sequence Generation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Bei Li, Chenglong Wang, Hang Zhou, Jingbo Zhu, Tongran Liu, Tong Xiao, Yifu Huo, Yimin Hu","submitted_at":"2023-08-04T09:35:45Z","abstract_excerpt":"Applying Reinforcement Learning (RL) to sequence generation models enables the direct optimization of long-term rewards (\\textit{e.g.,} BLEU and human feedback), but typically requires large-scale sampling over a space of action sequences. This is a computational challenge as presented by the practice of sequence generation problems, such as machine translation, where we often deal with a large action space (\\textit{e.g.,} a vocabulary) and a long action sequence (\\textit{e.g.,} a translation). In this work, we introduce two-stage sampling and dynamic sampling approaches to improve the samplin"},"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":"2308.02223","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-08-04T09:35:45Z","cross_cats_sorted":[],"title_canon_sha256":"af866bbf2a2a633146ad667bb67c293b3f22731604fbaabd42332b0caac4a746","abstract_canon_sha256":"7cf73833268a601c0c9ed87d423fc0ac14d32e9e50c17b7e3a1cf87578810955"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T06:37:39.487340Z","signature_b64":"3q6tCSCMrswgrK2hue+HYyvtkjNJTBXTNqUT7kybTmTT9a8nj39709aupZgBF765wCOWrCTZ1V8pRjey2YuIBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"258b61d9aaa81166351416a04a366b4427e7d274d326c1af4c6c4ec89e595855","last_reissued_at":"2026-07-05T06:37:39.486850Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T06:37:39.486850Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"ESRL: Efficient Sampling-based Reinforcement Learning for Sequence Generation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Bei Li, Chenglong Wang, Hang Zhou, Jingbo Zhu, Tongran Liu, Tong Xiao, Yifu Huo, Yimin Hu","submitted_at":"2023-08-04T09:35:45Z","abstract_excerpt":"Applying Reinforcement Learning (RL) to sequence generation models enables the direct optimization of long-term rewards (\\textit{e.g.,} BLEU and human feedback), but typically requires large-scale sampling over a space of action sequences. This is a computational challenge as presented by the practice of sequence generation problems, such as machine translation, where we often deal with a large action space (\\textit{e.g.,} a vocabulary) and a long action sequence (\\textit{e.g.,} a translation). In this work, we introduce two-stage sampling and dynamic sampling approaches to improve the samplin"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2308.02223","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/2308.02223/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":"2308.02223","created_at":"2026-07-05T06:37:39.486903+00:00"},{"alias_kind":"arxiv_version","alias_value":"2308.02223v1","created_at":"2026-07-05T06:37:39.486903+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2308.02223","created_at":"2026-07-05T06:37:39.486903+00:00"},{"alias_kind":"pith_short_12","alias_value":"EWFWDWNKVAIW","created_at":"2026-07-05T06:37:39.486903+00:00"},{"alias_kind":"pith_short_16","alias_value":"EWFWDWNKVAIWMNIU","created_at":"2026-07-05T06:37:39.486903+00:00"},{"alias_kind":"pith_short_8","alias_value":"EWFWDWNK","created_at":"2026-07-05T06:37:39.486903+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/EWFWDWNKVAIWMNIUC2QEUNTLIQ","json":"https://pith.science/pith/EWFWDWNKVAIWMNIUC2QEUNTLIQ.json","graph_json":"https://pith.science/api/pith-number/EWFWDWNKVAIWMNIUC2QEUNTLIQ/graph.json","events_json":"https://pith.science/api/pith-number/EWFWDWNKVAIWMNIUC2QEUNTLIQ/events.json","paper":"https://pith.science/paper/EWFWDWNK"},"agent_actions":{"view_html":"https://pith.science/pith/EWFWDWNKVAIWMNIUC2QEUNTLIQ","download_json":"https://pith.science/pith/EWFWDWNKVAIWMNIUC2QEUNTLIQ.json","view_paper":"https://pith.science/paper/EWFWDWNK","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2308.02223&json=true","fetch_graph":"https://pith.science/api/pith-number/EWFWDWNKVAIWMNIUC2QEUNTLIQ/graph.json","fetch_events":"https://pith.science/api/pith-number/EWFWDWNKVAIWMNIUC2QEUNTLIQ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/EWFWDWNKVAIWMNIUC2QEUNTLIQ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/EWFWDWNKVAIWMNIUC2QEUNTLIQ/action/storage_attestation","attest_author":"https://pith.science/pith/EWFWDWNKVAIWMNIUC2QEUNTLIQ/action/author_attestation","sign_citation":"https://pith.science/pith/EWFWDWNKVAIWMNIUC2QEUNTLIQ/action/citation_signature","submit_replication":"https://pith.science/pith/EWFWDWNKVAIWMNIUC2QEUNTLIQ/action/replication_record"}},"created_at":"2026-07-05T06:37:39.486903+00:00","updated_at":"2026-07-05T06:37:39.486903+00:00"}