{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2019:474AIVG2ZCPKI3TUHQNX2NNTZ5","short_pith_number":"pith:474AIVG2","schema_version":"1.0","canonical_sha256":"e7f80454dac89ea46e743c1b7d35b3cf7e05a5f337223e269666e62354253651","source":{"kind":"arxiv","id":"1909.01214","version":1},"attestation_state":"computed","paper":{"title":"Better Rewards Yield Better Summaries: Learning to Summarise Without References","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Christian M. Meyer, Florian B\\\"ohm, Ido Dagan, Iryna Gurevych, Ori Shapira, Yang Gao","submitted_at":"2019-09-03T14:30:13Z","abstract_excerpt":"Reinforcement Learning (RL) based document summarisation systems yield state-of-the-art performance in terms of ROUGE scores, because they directly use ROUGE as the rewards during training. However, summaries with high ROUGE scores often receive low human judgement. To find a better reward function that can guide RL to generate human-appealing summaries, we learn a reward function from human ratings on 2,500 summaries. Our reward function only takes the document and system summary as input. Hence, once trained, it can be used to train RL-based summarisation systems without using any reference "},"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":"1909.01214","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2019-09-03T14:30:13Z","cross_cats_sorted":[],"title_canon_sha256":"dab69358e30b4220084711aa4f14698a53a8eb2c35fc93fc51af9bb0db3994e2","abstract_canon_sha256":"23e52157f31d4977316df41cae1f256ae4ed3ea66e9b09508178d9e654cc2605"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T00:01:17.059395Z","signature_b64":"j29MfrdYF5HnNH4mO84vaHwvw2GgdYwjcO+rvAOAINbA48ovJ9I0jCasFaWeqAk29WSbxhAsqGiEydMnbVcuDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e7f80454dac89ea46e743c1b7d35b3cf7e05a5f337223e269666e62354253651","last_reissued_at":"2026-07-05T00:01:17.058939Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T00:01:17.058939Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Better Rewards Yield Better Summaries: Learning to Summarise Without References","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Christian M. Meyer, Florian B\\\"ohm, Ido Dagan, Iryna Gurevych, Ori Shapira, Yang Gao","submitted_at":"2019-09-03T14:30:13Z","abstract_excerpt":"Reinforcement Learning (RL) based document summarisation systems yield state-of-the-art performance in terms of ROUGE scores, because they directly use ROUGE as the rewards during training. However, summaries with high ROUGE scores often receive low human judgement. To find a better reward function that can guide RL to generate human-appealing summaries, we learn a reward function from human ratings on 2,500 summaries. Our reward function only takes the document and system summary as input. Hence, once trained, it can be used to train RL-based summarisation systems without using any reference "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1909.01214","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/1909.01214/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":"1909.01214","created_at":"2026-07-05T00:01:17.058996+00:00"},{"alias_kind":"arxiv_version","alias_value":"1909.01214v1","created_at":"2026-07-05T00:01:17.058996+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1909.01214","created_at":"2026-07-05T00:01:17.058996+00:00"},{"alias_kind":"pith_short_12","alias_value":"474AIVG2ZCPK","created_at":"2026-07-05T00:01:17.058996+00:00"},{"alias_kind":"pith_short_16","alias_value":"474AIVG2ZCPKI3TU","created_at":"2026-07-05T00:01:17.058996+00:00"},{"alias_kind":"pith_short_8","alias_value":"474AIVG2","created_at":"2026-07-05T00:01:17.058996+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2009.01325","citing_title":"Learning to summarize from human feedback","ref_index":3,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/474AIVG2ZCPKI3TUHQNX2NNTZ5","json":"https://pith.science/pith/474AIVG2ZCPKI3TUHQNX2NNTZ5.json","graph_json":"https://pith.science/api/pith-number/474AIVG2ZCPKI3TUHQNX2NNTZ5/graph.json","events_json":"https://pith.science/api/pith-number/474AIVG2ZCPKI3TUHQNX2NNTZ5/events.json","paper":"https://pith.science/paper/474AIVG2"},"agent_actions":{"view_html":"https://pith.science/pith/474AIVG2ZCPKI3TUHQNX2NNTZ5","download_json":"https://pith.science/pith/474AIVG2ZCPKI3TUHQNX2NNTZ5.json","view_paper":"https://pith.science/paper/474AIVG2","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1909.01214&json=true","fetch_graph":"https://pith.science/api/pith-number/474AIVG2ZCPKI3TUHQNX2NNTZ5/graph.json","fetch_events":"https://pith.science/api/pith-number/474AIVG2ZCPKI3TUHQNX2NNTZ5/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/474AIVG2ZCPKI3TUHQNX2NNTZ5/action/timestamp_anchor","attest_storage":"https://pith.science/pith/474AIVG2ZCPKI3TUHQNX2NNTZ5/action/storage_attestation","attest_author":"https://pith.science/pith/474AIVG2ZCPKI3TUHQNX2NNTZ5/action/author_attestation","sign_citation":"https://pith.science/pith/474AIVG2ZCPKI3TUHQNX2NNTZ5/action/citation_signature","submit_replication":"https://pith.science/pith/474AIVG2ZCPKI3TUHQNX2NNTZ5/action/replication_record"}},"created_at":"2026-07-05T00:01:17.058996+00:00","updated_at":"2026-07-05T00:01:17.058996+00:00"}