{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:ZOWZUQ3GKKO5SC5EOM7FAS7JFE","short_pith_number":"pith:ZOWZUQ3G","schema_version":"1.0","canonical_sha256":"cbad9a4366529dd90ba4733e504be92911b00ba76f4056237ad211c62880d643","source":{"kind":"arxiv","id":"2402.15610","version":2},"attestation_state":"computed","paper":{"title":"Selective \"Selective Prediction\": Reducing Unnecessary Abstention in Vision-Language Reasoning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Bill Yuchen Lin, Jack Hessel, Jesse Thomason, Khyathi Raghavi Chandu, Tanmay Gupta, Tejas Srinivasan, Yejin Choi","submitted_at":"2024-02-23T21:16:52Z","abstract_excerpt":"Selective prediction minimizes incorrect predictions from vision-language models (VLMs) by allowing them to abstain from answering when uncertain. However, when deploying a vision-language system with low tolerance for inaccurate predictions, selective prediction may be over-cautious and abstain too frequently, even on many correct predictions. We introduce ReCoVERR, an inference-time algorithm to reduce the over-abstention of a selective vision-language system without increasing the error rate of the system's predictions. When the VLM makes a low-confidence prediction, instead of abstaining R"},"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":"2402.15610","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-02-23T21:16:52Z","cross_cats_sorted":[],"title_canon_sha256":"0ee1e649bf3b81669ea48ec219643a0747be16751169147fd47375665f5e6a64","abstract_canon_sha256":"dd2f636c19f6d0d8118e50ccd4909fa453ef4e3b90b6442a8902c154af7c0917"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:31:12.232975Z","signature_b64":"BrxIAEX9+p+7sTzSzFBNDK5jPXElmFDAL1TGd3ORIZcpx0nMGkIC+S1gaD89K/aQw3gJC6Rt0BSoz9LrnVE4Cg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"cbad9a4366529dd90ba4733e504be92911b00ba76f4056237ad211c62880d643","last_reissued_at":"2026-07-05T08:31:12.232489Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:31:12.232489Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Selective \"Selective Prediction\": Reducing Unnecessary Abstention in Vision-Language Reasoning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Bill Yuchen Lin, Jack Hessel, Jesse Thomason, Khyathi Raghavi Chandu, Tanmay Gupta, Tejas Srinivasan, Yejin Choi","submitted_at":"2024-02-23T21:16:52Z","abstract_excerpt":"Selective prediction minimizes incorrect predictions from vision-language models (VLMs) by allowing them to abstain from answering when uncertain. However, when deploying a vision-language system with low tolerance for inaccurate predictions, selective prediction may be over-cautious and abstain too frequently, even on many correct predictions. We introduce ReCoVERR, an inference-time algorithm to reduce the over-abstention of a selective vision-language system without increasing the error rate of the system's predictions. When the VLM makes a low-confidence prediction, instead of abstaining R"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.15610","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/2402.15610/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":"2402.15610","created_at":"2026-07-05T08:31:12.232547+00:00"},{"alias_kind":"arxiv_version","alias_value":"2402.15610v2","created_at":"2026-07-05T08:31:12.232547+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.15610","created_at":"2026-07-05T08:31:12.232547+00:00"},{"alias_kind":"pith_short_12","alias_value":"ZOWZUQ3GKKO5","created_at":"2026-07-05T08:31:12.232547+00:00"},{"alias_kind":"pith_short_16","alias_value":"ZOWZUQ3GKKO5SC5E","created_at":"2026-07-05T08:31:12.232547+00:00"},{"alias_kind":"pith_short_8","alias_value":"ZOWZUQ3G","created_at":"2026-07-05T08:31:12.232547+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2603.19790","citing_title":"From Plausibility to Verifiability: Risk-Controlled Generative OCR with Vision-Language Models","ref_index":35,"is_internal_anchor":false},{"citing_arxiv_id":"2604.19895","citing_title":"Learning When Not to Decide: A Framework for Overcoming Factual Presumptuousness in AI Adjudication","ref_index":33,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/ZOWZUQ3GKKO5SC5EOM7FAS7JFE","json":"https://pith.science/pith/ZOWZUQ3GKKO5SC5EOM7FAS7JFE.json","graph_json":"https://pith.science/api/pith-number/ZOWZUQ3GKKO5SC5EOM7FAS7JFE/graph.json","events_json":"https://pith.science/api/pith-number/ZOWZUQ3GKKO5SC5EOM7FAS7JFE/events.json","paper":"https://pith.science/paper/ZOWZUQ3G"},"agent_actions":{"view_html":"https://pith.science/pith/ZOWZUQ3GKKO5SC5EOM7FAS7JFE","download_json":"https://pith.science/pith/ZOWZUQ3GKKO5SC5EOM7FAS7JFE.json","view_paper":"https://pith.science/paper/ZOWZUQ3G","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2402.15610&json=true","fetch_graph":"https://pith.science/api/pith-number/ZOWZUQ3GKKO5SC5EOM7FAS7JFE/graph.json","fetch_events":"https://pith.science/api/pith-number/ZOWZUQ3GKKO5SC5EOM7FAS7JFE/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/ZOWZUQ3GKKO5SC5EOM7FAS7JFE/action/timestamp_anchor","attest_storage":"https://pith.science/pith/ZOWZUQ3GKKO5SC5EOM7FAS7JFE/action/storage_attestation","attest_author":"https://pith.science/pith/ZOWZUQ3GKKO5SC5EOM7FAS7JFE/action/author_attestation","sign_citation":"https://pith.science/pith/ZOWZUQ3GKKO5SC5EOM7FAS7JFE/action/citation_signature","submit_replication":"https://pith.science/pith/ZOWZUQ3GKKO5SC5EOM7FAS7JFE/action/replication_record"}},"created_at":"2026-07-05T08:31:12.232547+00:00","updated_at":"2026-07-05T08:31:12.232547+00:00"}