{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2017:LMEEBE27K7TFDG6RUKBRM32UGC","short_pith_number":"pith:LMEEBE27","schema_version":"1.0","canonical_sha256":"5b0840935f57e6519bd1a283166f5430a8b2335822f5e1e52f120b69e2d1f67d","source":{"kind":"arxiv","id":"1707.06875","version":1},"attestation_state":"computed","paper":{"title":"Why We Need New Evaluation Metrics for NLG","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Amanda Cercas Curry, Jekaterina Novikova, Ond\\v{r}ej Du\\v{s}ek, Verena Rieser","submitted_at":"2017-07-21T12:47:03Z","abstract_excerpt":"The majority of NLG evaluation relies on automatic metrics, such as BLEU . In this paper, we motivate the need for novel, system- and data-independent automatic evaluation methods: We investigate a wide range of metrics, including state-of-the-art word-based and novel grammar-based ones, and demonstrate that they only weakly reflect human judgements of system outputs as generated by data-driven, end-to-end NLG. We also show that metric performance is data- and system-specific. Nevertheless, our results also suggest that automatic metrics perform reliably at system-level and can support system "},"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":"1707.06875","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2017-07-21T12:47:03Z","cross_cats_sorted":[],"title_canon_sha256":"2d00563cbedaeee32261adf6c4bb899ff307bdd9be7555fdaee9852920820d88","abstract_canon_sha256":"ff9c4e0bac48bf1b003765fb59544099914670fedbca441fbb296d4465858fcb"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T00:35:08.965626Z","signature_b64":"u+5NEch+Vlg/By2UtTIzKR7VaKf9SEvzmANVA8N3w+Qr1DLS0Kv+M+MDMjvhSSnlpJfwuaBCv7+YOVdKRtB3DQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"5b0840935f57e6519bd1a283166f5430a8b2335822f5e1e52f120b69e2d1f67d","last_reissued_at":"2026-05-18T00:35:08.965254Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T00:35:08.965254Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Why We Need New Evaluation Metrics for NLG","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Amanda Cercas Curry, Jekaterina Novikova, Ond\\v{r}ej Du\\v{s}ek, Verena Rieser","submitted_at":"2017-07-21T12:47:03Z","abstract_excerpt":"The majority of NLG evaluation relies on automatic metrics, such as BLEU . In this paper, we motivate the need for novel, system- and data-independent automatic evaluation methods: We investigate a wide range of metrics, including state-of-the-art word-based and novel grammar-based ones, and demonstrate that they only weakly reflect human judgements of system outputs as generated by data-driven, end-to-end NLG. We also show that metric performance is data- and system-specific. Nevertheless, our results also suggest that automatic metrics perform reliably at system-level and can support system "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1707.06875","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":"1707.06875","created_at":"2026-05-18T00:35:08.965323+00:00"},{"alias_kind":"arxiv_version","alias_value":"1707.06875v1","created_at":"2026-05-18T00:35:08.965323+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1707.06875","created_at":"2026-05-18T00:35:08.965323+00:00"},{"alias_kind":"pith_short_12","alias_value":"LMEEBE27K7TF","created_at":"2026-05-18T12:31:28.150371+00:00"},{"alias_kind":"pith_short_16","alias_value":"LMEEBE27K7TFDG6R","created_at":"2026-05-18T12:31:28.150371+00:00"},{"alias_kind":"pith_short_8","alias_value":"LMEEBE27","created_at":"2026-05-18T12:31:28.150371+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2605.19220","citing_title":"Position: Uncertainty Quantification in LLMs is Just Unsupervised Clustering","ref_index":40,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/LMEEBE27K7TFDG6RUKBRM32UGC","json":"https://pith.science/pith/LMEEBE27K7TFDG6RUKBRM32UGC.json","graph_json":"https://pith.science/api/pith-number/LMEEBE27K7TFDG6RUKBRM32UGC/graph.json","events_json":"https://pith.science/api/pith-number/LMEEBE27K7TFDG6RUKBRM32UGC/events.json","paper":"https://pith.science/paper/LMEEBE27"},"agent_actions":{"view_html":"https://pith.science/pith/LMEEBE27K7TFDG6RUKBRM32UGC","download_json":"https://pith.science/pith/LMEEBE27K7TFDG6RUKBRM32UGC.json","view_paper":"https://pith.science/paper/LMEEBE27","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1707.06875&json=true","fetch_graph":"https://pith.science/api/pith-number/LMEEBE27K7TFDG6RUKBRM32UGC/graph.json","fetch_events":"https://pith.science/api/pith-number/LMEEBE27K7TFDG6RUKBRM32UGC/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/LMEEBE27K7TFDG6RUKBRM32UGC/action/timestamp_anchor","attest_storage":"https://pith.science/pith/LMEEBE27K7TFDG6RUKBRM32UGC/action/storage_attestation","attest_author":"https://pith.science/pith/LMEEBE27K7TFDG6RUKBRM32UGC/action/author_attestation","sign_citation":"https://pith.science/pith/LMEEBE27K7TFDG6RUKBRM32UGC/action/citation_signature","submit_replication":"https://pith.science/pith/LMEEBE27K7TFDG6RUKBRM32UGC/action/replication_record"}},"created_at":"2026-05-18T00:35:08.965323+00:00","updated_at":"2026-05-18T00:35:08.965323+00:00"}