{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:HPMHKLMGFV3OISLBJSW4RW5IEC","short_pith_number":"pith:HPMHKLMG","schema_version":"1.0","canonical_sha256":"3bd8752d862d76e449614cadc8dba8208438705f02761bb7b707602b0c8f7818","source":{"kind":"arxiv","id":"2203.06063","version":2},"attestation_state":"computed","paper":{"title":"Active Evaluation: Efficient NLG Evaluation with Few Pairwise Comparisons","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Akash Kumar Mohankumar, Mitesh M. Khapra","submitted_at":"2022-03-11T16:39:15Z","abstract_excerpt":"Recent studies have shown the advantages of evaluating NLG systems using pairwise comparisons as opposed to direct assessment. Given $k$ systems, a naive approach for identifying the top-ranked system would be to uniformly obtain pairwise comparisons from all ${k \\choose 2}$ pairs of systems. However, this can be very expensive as the number of human annotations required would grow quadratically with $k$. In this work, we introduce Active Evaluation, a framework to efficiently identify the top-ranked system by actively choosing system pairs for comparison using dueling bandit algorithms. We pe"},"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":"2203.06063","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2022-03-11T16:39:15Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"1540ddbd55e5f7b7cd9b7a16d591d4651352f9b6025d1a46616bedfc8a39709e","abstract_canon_sha256":"aefdf90388bdbd8c2f35188ddbeba8ab82926df1ea0ed851409a4b379151b184"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T04:15:14.812793Z","signature_b64":"ugqWMtO9wT/rROL/z0CrUscXK4m5bQ77gfdIn8aS9Bxzb8JJXxIQVq32Z03AtHDBMPqiixf0mActkheTnWOlBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"3bd8752d862d76e449614cadc8dba8208438705f02761bb7b707602b0c8f7818","last_reissued_at":"2026-07-05T04:15:14.812336Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T04:15:14.812336Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Active Evaluation: Efficient NLG Evaluation with Few Pairwise Comparisons","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Akash Kumar Mohankumar, Mitesh M. Khapra","submitted_at":"2022-03-11T16:39:15Z","abstract_excerpt":"Recent studies have shown the advantages of evaluating NLG systems using pairwise comparisons as opposed to direct assessment. Given $k$ systems, a naive approach for identifying the top-ranked system would be to uniformly obtain pairwise comparisons from all ${k \\choose 2}$ pairs of systems. However, this can be very expensive as the number of human annotations required would grow quadratically with $k$. In this work, we introduce Active Evaluation, a framework to efficiently identify the top-ranked system by actively choosing system pairs for comparison using dueling bandit algorithms. We pe"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2203.06063","kind":"arxiv","version":2},"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/2203.06063/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":"2203.06063","created_at":"2026-07-05T04:15:14.812395+00:00"},{"alias_kind":"arxiv_version","alias_value":"2203.06063v2","created_at":"2026-07-05T04:15:14.812395+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2203.06063","created_at":"2026-07-05T04:15:14.812395+00:00"},{"alias_kind":"pith_short_12","alias_value":"HPMHKLMGFV3O","created_at":"2026-07-05T04:15:14.812395+00:00"},{"alias_kind":"pith_short_16","alias_value":"HPMHKLMGFV3OISLB","created_at":"2026-07-05T04:15:14.812395+00:00"},{"alias_kind":"pith_short_8","alias_value":"HPMHKLMG","created_at":"2026-07-05T04:15:14.812395+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/HPMHKLMGFV3OISLBJSW4RW5IEC","json":"https://pith.science/pith/HPMHKLMGFV3OISLBJSW4RW5IEC.json","graph_json":"https://pith.science/api/pith-number/HPMHKLMGFV3OISLBJSW4RW5IEC/graph.json","events_json":"https://pith.science/api/pith-number/HPMHKLMGFV3OISLBJSW4RW5IEC/events.json","paper":"https://pith.science/paper/HPMHKLMG"},"agent_actions":{"view_html":"https://pith.science/pith/HPMHKLMGFV3OISLBJSW4RW5IEC","download_json":"https://pith.science/pith/HPMHKLMGFV3OISLBJSW4RW5IEC.json","view_paper":"https://pith.science/paper/HPMHKLMG","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2203.06063&json=true","fetch_graph":"https://pith.science/api/pith-number/HPMHKLMGFV3OISLBJSW4RW5IEC/graph.json","fetch_events":"https://pith.science/api/pith-number/HPMHKLMGFV3OISLBJSW4RW5IEC/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/HPMHKLMGFV3OISLBJSW4RW5IEC/action/timestamp_anchor","attest_storage":"https://pith.science/pith/HPMHKLMGFV3OISLBJSW4RW5IEC/action/storage_attestation","attest_author":"https://pith.science/pith/HPMHKLMGFV3OISLBJSW4RW5IEC/action/author_attestation","sign_citation":"https://pith.science/pith/HPMHKLMGFV3OISLBJSW4RW5IEC/action/citation_signature","submit_replication":"https://pith.science/pith/HPMHKLMGFV3OISLBJSW4RW5IEC/action/replication_record"}},"created_at":"2026-07-05T04:15:14.812395+00:00","updated_at":"2026-07-05T04:15:14.812395+00:00"}