{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2016:EZT5KX4H23E6SIYHTBLJDBAAVI","short_pith_number":"pith:EZT5KX4H","schema_version":"1.0","canonical_sha256":"2667d55f87d6c9e923079856918400aa3da7c303cb5c54db3b1e36a1ab5736f0","source":{"kind":"arxiv","id":"1611.02796","version":9},"attestation_state":"computed","paper":{"title":"Sequence Tutor: Conservative Fine-Tuning of Sequence Generation Models with KL-control","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Douglas Eck, Dzmitry Bahdanau, Jos\\'e Miguel Hern\\'andez-Lobato, Natasha Jaques, Richard E. Turner, Shixiang Gu","submitted_at":"2016-11-09T01:46:32Z","abstract_excerpt":"This paper proposes a general method for improving the structure and quality of sequences generated by a recurrent neural network (RNN), while maintaining information originally learned from data, as well as sample diversity. An RNN is first pre-trained on data using maximum likelihood estimation (MLE), and the probability distribution over the next token in the sequence learned by this model is treated as a prior policy. Another RNN is then trained using reinforcement learning (RL) to generate higher-quality outputs that account for domain-specific incentives while retaining proximity to the "},"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":"1611.02796","kind":"arxiv","version":9},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2016-11-09T01:46:32Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"056ae21c44a8f8bf0bee5126878617ccfb343a168f6880258e62c43c173f521c","abstract_canon_sha256":"d4e340dc4e9b97cdfaffca47c670812b32e0d337c190193269f77730880cb832"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T00:32:44.214013Z","signature_b64":"lha6Upm47cSWVN9cVLq6TEcArwu2FweI/UUkQeQHu4kWgjuVDnA1v89fzUkxIkFiC536qLi3e2nia9YG8W2KCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"2667d55f87d6c9e923079856918400aa3da7c303cb5c54db3b1e36a1ab5736f0","last_reissued_at":"2026-05-18T00:32:44.213393Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T00:32:44.213393Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Sequence Tutor: Conservative Fine-Tuning of Sequence Generation Models with KL-control","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Douglas Eck, Dzmitry Bahdanau, Jos\\'e Miguel Hern\\'andez-Lobato, Natasha Jaques, Richard E. Turner, Shixiang Gu","submitted_at":"2016-11-09T01:46:32Z","abstract_excerpt":"This paper proposes a general method for improving the structure and quality of sequences generated by a recurrent neural network (RNN), while maintaining information originally learned from data, as well as sample diversity. An RNN is first pre-trained on data using maximum likelihood estimation (MLE), and the probability distribution over the next token in the sequence learned by this model is treated as a prior policy. Another RNN is then trained using reinforcement learning (RL) to generate higher-quality outputs that account for domain-specific incentives while retaining proximity to the "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1611.02796","kind":"arxiv","version":9},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"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":"1611.02796","created_at":"2026-05-18T00:32:44.213479+00:00"},{"alias_kind":"arxiv_version","alias_value":"1611.02796v9","created_at":"2026-05-18T00:32:44.213479+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1611.02796","created_at":"2026-05-18T00:32:44.213479+00:00"},{"alias_kind":"pith_short_12","alias_value":"EZT5KX4H23E6","created_at":"2026-05-18T12:30:15.759754+00:00"},{"alias_kind":"pith_short_16","alias_value":"EZT5KX4H23E6SIYH","created_at":"2026-05-18T12:30:15.759754+00:00"},{"alias_kind":"pith_short_8","alias_value":"EZT5KX4H","created_at":"2026-05-18T12:30:15.759754+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/EZT5KX4H23E6SIYHTBLJDBAAVI","json":"https://pith.science/pith/EZT5KX4H23E6SIYHTBLJDBAAVI.json","graph_json":"https://pith.science/api/pith-number/EZT5KX4H23E6SIYHTBLJDBAAVI/graph.json","events_json":"https://pith.science/api/pith-number/EZT5KX4H23E6SIYHTBLJDBAAVI/events.json","paper":"https://pith.science/paper/EZT5KX4H"},"agent_actions":{"view_html":"https://pith.science/pith/EZT5KX4H23E6SIYHTBLJDBAAVI","download_json":"https://pith.science/pith/EZT5KX4H23E6SIYHTBLJDBAAVI.json","view_paper":"https://pith.science/paper/EZT5KX4H","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1611.02796&json=true","fetch_graph":"https://pith.science/api/pith-number/EZT5KX4H23E6SIYHTBLJDBAAVI/graph.json","fetch_events":"https://pith.science/api/pith-number/EZT5KX4H23E6SIYHTBLJDBAAVI/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/EZT5KX4H23E6SIYHTBLJDBAAVI/action/timestamp_anchor","attest_storage":"https://pith.science/pith/EZT5KX4H23E6SIYHTBLJDBAAVI/action/storage_attestation","attest_author":"https://pith.science/pith/EZT5KX4H23E6SIYHTBLJDBAAVI/action/author_attestation","sign_citation":"https://pith.science/pith/EZT5KX4H23E6SIYHTBLJDBAAVI/action/citation_signature","submit_replication":"https://pith.science/pith/EZT5KX4H23E6SIYHTBLJDBAAVI/action/replication_record"}},"created_at":"2026-05-18T00:32:44.213479+00:00","updated_at":"2026-05-18T00:32:44.213479+00:00"}