{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:AAOFZC6IGQESA7V5OAAGBRHCQ6","short_pith_number":"pith:AAOFZC6I","schema_version":"1.0","canonical_sha256":"001c5c8bc83409207ebd700060c4e28796f232b2302366281c9c85147ba795bc","source":{"kind":"arxiv","id":"2310.14670","version":2},"attestation_state":"computed","paper":{"title":"Dataset Bias Mitigation in Multiple-Choice Visual Question Answering and Beyond","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL","cs.LG","cs.MM"],"primary_cat":"cs.CV","authors_text":"Haoxuan You, Kai-Wei Chang, Keyang Xu, Long Chen, Noel Codella, Shih-Fu Chang, Wenhao Li, Yicheng He, Zhecan Wang","submitted_at":"2023-10-23T08:09:42Z","abstract_excerpt":"Vision-language (VL) understanding tasks evaluate models' comprehension of complex visual scenes through multiple-choice questions. However, we have identified two dataset biases that models can exploit as shortcuts to resolve various VL tasks correctly without proper understanding. The first type of dataset bias is \\emph{Unbalanced Matching} bias, where the correct answer overlaps the question and image more than the incorrect answers. The second type of dataset bias is \\emph{Distractor Similarity} bias, where incorrect answers are overly dissimilar to the correct answer but significantly sim"},"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":"2310.14670","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2023-10-23T08:09:42Z","cross_cats_sorted":["cs.AI","cs.CL","cs.LG","cs.MM"],"title_canon_sha256":"c55e9655c081d5dfd2835651e0f59a47f856930db9221c5386fb8be8d6d9554d","abstract_canon_sha256":"f6df2c053bc627e97bd1377ead413e17ce099d6b5b4f188c1bf9f964109ef0b9"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:07:40.097909Z","signature_b64":"RDfhuAN7FzEo6RGI+FunoyzCuzFzh2g0Zs7oVVcN1GuHYDZSQsNjLPuJiYyjc72cg2biZ+cDqJzheRK/lspRBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"001c5c8bc83409207ebd700060c4e28796f232b2302366281c9c85147ba795bc","last_reissued_at":"2026-07-05T07:07:40.097447Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:07:40.097447Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Dataset Bias Mitigation in Multiple-Choice Visual Question Answering and Beyond","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL","cs.LG","cs.MM"],"primary_cat":"cs.CV","authors_text":"Haoxuan You, Kai-Wei Chang, Keyang Xu, Long Chen, Noel Codella, Shih-Fu Chang, Wenhao Li, Yicheng He, Zhecan Wang","submitted_at":"2023-10-23T08:09:42Z","abstract_excerpt":"Vision-language (VL) understanding tasks evaluate models' comprehension of complex visual scenes through multiple-choice questions. However, we have identified two dataset biases that models can exploit as shortcuts to resolve various VL tasks correctly without proper understanding. The first type of dataset bias is \\emph{Unbalanced Matching} bias, where the correct answer overlaps the question and image more than the incorrect answers. The second type of dataset bias is \\emph{Distractor Similarity} bias, where incorrect answers are overly dissimilar to the correct answer but significantly sim"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2310.14670","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/2310.14670/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":"2310.14670","created_at":"2026-07-05T07:07:40.097508+00:00"},{"alias_kind":"arxiv_version","alias_value":"2310.14670v2","created_at":"2026-07-05T07:07:40.097508+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2310.14670","created_at":"2026-07-05T07:07:40.097508+00:00"},{"alias_kind":"pith_short_12","alias_value":"AAOFZC6IGQES","created_at":"2026-07-05T07:07:40.097508+00:00"},{"alias_kind":"pith_short_16","alias_value":"AAOFZC6IGQESA7V5","created_at":"2026-07-05T07:07:40.097508+00:00"},{"alias_kind":"pith_short_8","alias_value":"AAOFZC6I","created_at":"2026-07-05T07:07:40.097508+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2508.13428","citing_title":"Mitigating Easy Option Bias in Multiple-Choice Question Answering","ref_index":48,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/AAOFZC6IGQESA7V5OAAGBRHCQ6","json":"https://pith.science/pith/AAOFZC6IGQESA7V5OAAGBRHCQ6.json","graph_json":"https://pith.science/api/pith-number/AAOFZC6IGQESA7V5OAAGBRHCQ6/graph.json","events_json":"https://pith.science/api/pith-number/AAOFZC6IGQESA7V5OAAGBRHCQ6/events.json","paper":"https://pith.science/paper/AAOFZC6I"},"agent_actions":{"view_html":"https://pith.science/pith/AAOFZC6IGQESA7V5OAAGBRHCQ6","download_json":"https://pith.science/pith/AAOFZC6IGQESA7V5OAAGBRHCQ6.json","view_paper":"https://pith.science/paper/AAOFZC6I","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2310.14670&json=true","fetch_graph":"https://pith.science/api/pith-number/AAOFZC6IGQESA7V5OAAGBRHCQ6/graph.json","fetch_events":"https://pith.science/api/pith-number/AAOFZC6IGQESA7V5OAAGBRHCQ6/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/AAOFZC6IGQESA7V5OAAGBRHCQ6/action/timestamp_anchor","attest_storage":"https://pith.science/pith/AAOFZC6IGQESA7V5OAAGBRHCQ6/action/storage_attestation","attest_author":"https://pith.science/pith/AAOFZC6IGQESA7V5OAAGBRHCQ6/action/author_attestation","sign_citation":"https://pith.science/pith/AAOFZC6IGQESA7V5OAAGBRHCQ6/action/citation_signature","submit_replication":"https://pith.science/pith/AAOFZC6IGQESA7V5OAAGBRHCQ6/action/replication_record"}},"created_at":"2026-07-05T07:07:40.097508+00:00","updated_at":"2026-07-05T07:07:40.097508+00:00"}