{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2018:QZING43BF4KAUW2FOBRBLPDBQG","short_pith_number":"pith:QZING43B","schema_version":"1.0","canonical_sha256":"8650d373612f140a5b45706215bc6181a6caecec58fea36851572aad96cf40a5","source":{"kind":"arxiv","id":"1805.03766","version":1},"attestation_state":"computed","paper":{"title":"Discourse-Aware Neural Rewards for Coherent Text Generation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Antoine Bosselut, Asli Celikyilmaz, Jianfeng Gao, Po-Sen Huang, Xiaodong He, Yejin Choi","submitted_at":"2018-05-10T00:51:06Z","abstract_excerpt":"In this paper, we investigate the use of discourse-aware rewards with reinforcement learning to guide a model to generate long, coherent text. In particular, we propose to learn neural rewards to model cross-sentence ordering as a means to approximate desired discourse structure. Empirical results demonstrate that a generator trained with the learned reward produces more coherent and less repetitive text than models trained with cross-entropy or with reinforcement learning with commonly used scores as rewards."},"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":"1805.03766","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2018-05-10T00:51:06Z","cross_cats_sorted":[],"title_canon_sha256":"b27741d2db3ab81b70c13affb96ac61dde4fbfdf58b663a5ad064d262ee05f7b","abstract_canon_sha256":"a0f874233194917646243a77bf8ecee95267281860ea84b1f436609cafbc9d7f"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T00:16:18.829157Z","signature_b64":"VIKlMMdSQssg9L2fSGdTlnPwitXzeB/KKcXOWY/Ss0UfAqpJhVUpyCub813X7kc04iK1SnzMLSeFqcvitXYFAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"8650d373612f140a5b45706215bc6181a6caecec58fea36851572aad96cf40a5","last_reissued_at":"2026-05-18T00:16:18.828643Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T00:16:18.828643Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Discourse-Aware Neural Rewards for Coherent Text Generation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Antoine Bosselut, Asli Celikyilmaz, Jianfeng Gao, Po-Sen Huang, Xiaodong He, Yejin Choi","submitted_at":"2018-05-10T00:51:06Z","abstract_excerpt":"In this paper, we investigate the use of discourse-aware rewards with reinforcement learning to guide a model to generate long, coherent text. In particular, we propose to learn neural rewards to model cross-sentence ordering as a means to approximate desired discourse structure. Empirical results demonstrate that a generator trained with the learned reward produces more coherent and less repetitive text than models trained with cross-entropy or with reinforcement learning with commonly used scores as rewards."},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1805.03766","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":""},"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":"1805.03766","created_at":"2026-05-18T00:16:18.828732+00:00"},{"alias_kind":"arxiv_version","alias_value":"1805.03766v1","created_at":"2026-05-18T00:16:18.828732+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1805.03766","created_at":"2026-05-18T00:16:18.828732+00:00"},{"alias_kind":"pith_short_12","alias_value":"QZING43BF4KA","created_at":"2026-05-18T12:32:50.500415+00:00"},{"alias_kind":"pith_short_16","alias_value":"QZING43BF4KAUW2F","created_at":"2026-05-18T12:32:50.500415+00:00"},{"alias_kind":"pith_short_8","alias_value":"QZING43B","created_at":"2026-05-18T12:32:50.500415+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"1909.05364","citing_title":"TransSent: Towards Generation of Structured Sentences with Discourse Marker","ref_index":2013,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/QZING43BF4KAUW2FOBRBLPDBQG","json":"https://pith.science/pith/QZING43BF4KAUW2FOBRBLPDBQG.json","graph_json":"https://pith.science/api/pith-number/QZING43BF4KAUW2FOBRBLPDBQG/graph.json","events_json":"https://pith.science/api/pith-number/QZING43BF4KAUW2FOBRBLPDBQG/events.json","paper":"https://pith.science/paper/QZING43B"},"agent_actions":{"view_html":"https://pith.science/pith/QZING43BF4KAUW2FOBRBLPDBQG","download_json":"https://pith.science/pith/QZING43BF4KAUW2FOBRBLPDBQG.json","view_paper":"https://pith.science/paper/QZING43B","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1805.03766&json=true","fetch_graph":"https://pith.science/api/pith-number/QZING43BF4KAUW2FOBRBLPDBQG/graph.json","fetch_events":"https://pith.science/api/pith-number/QZING43BF4KAUW2FOBRBLPDBQG/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/QZING43BF4KAUW2FOBRBLPDBQG/action/timestamp_anchor","attest_storage":"https://pith.science/pith/QZING43BF4KAUW2FOBRBLPDBQG/action/storage_attestation","attest_author":"https://pith.science/pith/QZING43BF4KAUW2FOBRBLPDBQG/action/author_attestation","sign_citation":"https://pith.science/pith/QZING43BF4KAUW2FOBRBLPDBQG/action/citation_signature","submit_replication":"https://pith.science/pith/QZING43BF4KAUW2FOBRBLPDBQG/action/replication_record"}},"created_at":"2026-05-18T00:16:18.828732+00:00","updated_at":"2026-05-18T00:16:18.828732+00:00"}