{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:AKMZJYJF36DCRZOVGRZLK7WR2S","short_pith_number":"pith:AKMZJYJF","schema_version":"1.0","canonical_sha256":"029994e125df8628e5d53472b57ed1d4b0b3997017f6f96d8e7817c890e316ca","source":{"kind":"arxiv","id":"2407.07840","version":3},"attestation_state":"computed","paper":{"title":"Decompose and Compare Consistency: Measuring VLMs' Answer Reliability via Task-Decomposition Consistency Comparison","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.CV","authors_text":"Aishwarya Agrawal, Qian Yang, Weixiang Yan","submitted_at":"2024-07-10T17:00:29Z","abstract_excerpt":"Despite tremendous advancements, current state-of-the-art Vision-Language Models (VLMs) are still far from perfect. They tend to hallucinate and may generate biased responses. In such circumstances, having a way to assess the reliability of a given response generated by a VLM is quite useful. Existing methods, such as estimating uncertainty using answer likelihoods or prompt-based confidence generation, often suffer from overconfidence. Other methods use self-consistency comparison but are affected by confirmation biases. To alleviate these, we propose Decompose and Compare Consistency (DeCC) "},"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":"2407.07840","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2024-07-10T17:00:29Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"b8c60bdbb3b0354879f98eb50c732eb9a7f96e37377d2f03e970029b4ba14a64","abstract_canon_sha256":"da3369b2c41a26d1fc56796c6afc4661487ef47657e57f4b45abf807f6aab2a2"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:17:45.708657Z","signature_b64":"4LBrd0clf8RAVyOwQV2ArXIB9U0odKMzhpAsgHmqmBn8NmmHO4mMn35CWEl8swmxP9mZqnfJHn/btWE0wnGCAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"029994e125df8628e5d53472b57ed1d4b0b3997017f6f96d8e7817c890e316ca","last_reissued_at":"2026-07-05T09:17:45.708118Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:17:45.708118Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Decompose and Compare Consistency: Measuring VLMs' Answer Reliability via Task-Decomposition Consistency Comparison","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.CV","authors_text":"Aishwarya Agrawal, Qian Yang, Weixiang Yan","submitted_at":"2024-07-10T17:00:29Z","abstract_excerpt":"Despite tremendous advancements, current state-of-the-art Vision-Language Models (VLMs) are still far from perfect. They tend to hallucinate and may generate biased responses. In such circumstances, having a way to assess the reliability of a given response generated by a VLM is quite useful. Existing methods, such as estimating uncertainty using answer likelihoods or prompt-based confidence generation, often suffer from overconfidence. Other methods use self-consistency comparison but are affected by confirmation biases. To alleviate these, we propose Decompose and Compare Consistency (DeCC) "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2407.07840","kind":"arxiv","version":3},"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/2407.07840/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":"2407.07840","created_at":"2026-07-05T09:17:45.708191+00:00"},{"alias_kind":"arxiv_version","alias_value":"2407.07840v3","created_at":"2026-07-05T09:17:45.708191+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2407.07840","created_at":"2026-07-05T09:17:45.708191+00:00"},{"alias_kind":"pith_short_12","alias_value":"AKMZJYJF36DC","created_at":"2026-07-05T09:17:45.708191+00:00"},{"alias_kind":"pith_short_16","alias_value":"AKMZJYJF36DCRZOV","created_at":"2026-07-05T09:17:45.708191+00:00"},{"alias_kind":"pith_short_8","alias_value":"AKMZJYJF","created_at":"2026-07-05T09:17:45.708191+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2608.05675","citing_title":"Consistency Has a Computable Blind Spot: A Commutation Theory of Label-Free Reliability for Vision-Language Figure Reading","ref_index":2,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/AKMZJYJF36DCRZOVGRZLK7WR2S","json":"https://pith.science/pith/AKMZJYJF36DCRZOVGRZLK7WR2S.json","graph_json":"https://pith.science/api/pith-number/AKMZJYJF36DCRZOVGRZLK7WR2S/graph.json","events_json":"https://pith.science/api/pith-number/AKMZJYJF36DCRZOVGRZLK7WR2S/events.json","paper":"https://pith.science/paper/AKMZJYJF"},"agent_actions":{"view_html":"https://pith.science/pith/AKMZJYJF36DCRZOVGRZLK7WR2S","download_json":"https://pith.science/pith/AKMZJYJF36DCRZOVGRZLK7WR2S.json","view_paper":"https://pith.science/paper/AKMZJYJF","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2407.07840&json=true","fetch_graph":"https://pith.science/api/pith-number/AKMZJYJF36DCRZOVGRZLK7WR2S/graph.json","fetch_events":"https://pith.science/api/pith-number/AKMZJYJF36DCRZOVGRZLK7WR2S/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/AKMZJYJF36DCRZOVGRZLK7WR2S/action/timestamp_anchor","attest_storage":"https://pith.science/pith/AKMZJYJF36DCRZOVGRZLK7WR2S/action/storage_attestation","attest_author":"https://pith.science/pith/AKMZJYJF36DCRZOVGRZLK7WR2S/action/author_attestation","sign_citation":"https://pith.science/pith/AKMZJYJF36DCRZOVGRZLK7WR2S/action/citation_signature","submit_replication":"https://pith.science/pith/AKMZJYJF36DCRZOVGRZLK7WR2S/action/replication_record"}},"created_at":"2026-07-05T09:17:45.708191+00:00","updated_at":"2026-07-05T09:17:45.708191+00:00"}