{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:Y5H2EQDT2ZRYLQ7UEUUIH3DU6W","short_pith_number":"pith:Y5H2EQDT","schema_version":"1.0","canonical_sha256":"c74fa24073d66385c3f4252883ec74f593b0f0b53efc5691d614a230fa9c3301","source":{"kind":"arxiv","id":"2310.00410","version":1},"attestation_state":"computed","paper":{"title":"Open-Domain Dialogue Quality Evaluation: Deriving Nugget-level Scores from Turn-level Scores","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.IR"],"primary_cat":"cs.CL","authors_text":"Akihisa Watanabe, Rikiya Takehi, Tetsuya Sakai","submitted_at":"2023-09-30T15:14:50Z","abstract_excerpt":"Existing dialogue quality evaluation systems can return a score for a given system turn from a particular viewpoint, e.g., engagingness. However, to improve dialogue systems by locating exactly where in a system turn potential problems lie, a more fine-grained evaluation may be necessary. We therefore propose an evaluation approach where a turn is decomposed into nuggets (i.e., expressions associated with a dialogue act), and nugget-level evaluation is enabled by leveraging an existing turn-level evaluation system. We demonstrate the potential effectiveness of our evaluation method through a c"},"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":"2310.00410","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-09-30T15:14:50Z","cross_cats_sorted":["cs.IR"],"title_canon_sha256":"e9256d294be625f29ebc482f2cc69efb99d6dd8778d676fe2bde7429a610d685","abstract_canon_sha256":"cdc281fa0fd9ac089691f5ea62d51cc8007c0c7a5118d55c1a8c940471214b5c"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T06:56:05.362089Z","signature_b64":"bruXbsa/5vQ8+/l9/QDDjup184fVFRHUe1v/okc1jT9Kg4FBvEmE/gVff7R9lKzEp++MTH5AwkNfZbnRvr0/Dw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c74fa24073d66385c3f4252883ec74f593b0f0b53efc5691d614a230fa9c3301","last_reissued_at":"2026-07-05T06:56:05.361579Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T06:56:05.361579Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Open-Domain Dialogue Quality Evaluation: Deriving Nugget-level Scores from Turn-level Scores","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.IR"],"primary_cat":"cs.CL","authors_text":"Akihisa Watanabe, Rikiya Takehi, Tetsuya Sakai","submitted_at":"2023-09-30T15:14:50Z","abstract_excerpt":"Existing dialogue quality evaluation systems can return a score for a given system turn from a particular viewpoint, e.g., engagingness. However, to improve dialogue systems by locating exactly where in a system turn potential problems lie, a more fine-grained evaluation may be necessary. We therefore propose an evaluation approach where a turn is decomposed into nuggets (i.e., expressions associated with a dialogue act), and nugget-level evaluation is enabled by leveraging an existing turn-level evaluation system. We demonstrate the potential effectiveness of our evaluation method through a c"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2310.00410","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/2310.00410/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":"2310.00410","created_at":"2026-07-05T06:56:05.361641+00:00"},{"alias_kind":"arxiv_version","alias_value":"2310.00410v1","created_at":"2026-07-05T06:56:05.361641+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2310.00410","created_at":"2026-07-05T06:56:05.361641+00:00"},{"alias_kind":"pith_short_12","alias_value":"Y5H2EQDT2ZRY","created_at":"2026-07-05T06:56:05.361641+00:00"},{"alias_kind":"pith_short_16","alias_value":"Y5H2EQDT2ZRYLQ7U","created_at":"2026-07-05T06:56:05.361641+00:00"},{"alias_kind":"pith_short_8","alias_value":"Y5H2EQDT","created_at":"2026-07-05T06:56:05.361641+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/Y5H2EQDT2ZRYLQ7UEUUIH3DU6W","json":"https://pith.science/pith/Y5H2EQDT2ZRYLQ7UEUUIH3DU6W.json","graph_json":"https://pith.science/api/pith-number/Y5H2EQDT2ZRYLQ7UEUUIH3DU6W/graph.json","events_json":"https://pith.science/api/pith-number/Y5H2EQDT2ZRYLQ7UEUUIH3DU6W/events.json","paper":"https://pith.science/paper/Y5H2EQDT"},"agent_actions":{"view_html":"https://pith.science/pith/Y5H2EQDT2ZRYLQ7UEUUIH3DU6W","download_json":"https://pith.science/pith/Y5H2EQDT2ZRYLQ7UEUUIH3DU6W.json","view_paper":"https://pith.science/paper/Y5H2EQDT","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2310.00410&json=true","fetch_graph":"https://pith.science/api/pith-number/Y5H2EQDT2ZRYLQ7UEUUIH3DU6W/graph.json","fetch_events":"https://pith.science/api/pith-number/Y5H2EQDT2ZRYLQ7UEUUIH3DU6W/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/Y5H2EQDT2ZRYLQ7UEUUIH3DU6W/action/timestamp_anchor","attest_storage":"https://pith.science/pith/Y5H2EQDT2ZRYLQ7UEUUIH3DU6W/action/storage_attestation","attest_author":"https://pith.science/pith/Y5H2EQDT2ZRYLQ7UEUUIH3DU6W/action/author_attestation","sign_citation":"https://pith.science/pith/Y5H2EQDT2ZRYLQ7UEUUIH3DU6W/action/citation_signature","submit_replication":"https://pith.science/pith/Y5H2EQDT2ZRYLQ7UEUUIH3DU6W/action/replication_record"}},"created_at":"2026-07-05T06:56:05.361641+00:00","updated_at":"2026-07-05T06:56:05.361641+00:00"}