{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:22D77WXEKTPAJDZTVLJPNVQPYN","short_pith_number":"pith:22D77WXE","schema_version":"1.0","canonical_sha256":"d687ffdae454de048f33aad2f6d60fc3585de2acfba45268cac4dbd390618208","source":{"kind":"arxiv","id":"2504.11038","version":1},"attestation_state":"computed","paper":{"title":"QAVA: Query-Agnostic Visual Attack to Large Vision-Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Jiansheng Chen, Ruobing Xie, Xingwu Sun, Yudong Zhang, Yu Wang, Zhanhui Kang","submitted_at":"2025-04-15T10:00:01Z","abstract_excerpt":"In typical multimodal tasks, such as Visual Question Answering (VQA), adversarial attacks targeting a specific image and question can lead large vision-language models (LVLMs) to provide incorrect answers. However, it is common for a single image to be associated with multiple questions, and LVLMs may still answer other questions correctly even for an adversarial image attacked by a specific question. To address this, we introduce the query-agnostic visual attack (QAVA), which aims to create robust adversarial examples that generate incorrect responses to unspecified and unknown questions. Com"},"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":"2504.11038","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2025-04-15T10:00:01Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"8607b9ad8748e6fdbdd4432959853504f8784e4807545fbfa5464c7e1b6e1295","abstract_canon_sha256":"9d2f9bc24460268e5370396a41cb3c68d18c46aef71bb6ee63596dc68716dd81"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:49:28.149731Z","signature_b64":"qYHVg5WW+mkqVCaJELdRWk24r9HxirXWucXnIgDqKauumedJfrwuDSCIqFot8yd0ykpduf/cmuWTfLBL1vMWBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"d687ffdae454de048f33aad2f6d60fc3585de2acfba45268cac4dbd390618208","last_reissued_at":"2026-07-05T10:49:28.149299Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:49:28.149299Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"QAVA: Query-Agnostic Visual Attack to Large Vision-Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Jiansheng Chen, Ruobing Xie, Xingwu Sun, Yudong Zhang, Yu Wang, Zhanhui Kang","submitted_at":"2025-04-15T10:00:01Z","abstract_excerpt":"In typical multimodal tasks, such as Visual Question Answering (VQA), adversarial attacks targeting a specific image and question can lead large vision-language models (LVLMs) to provide incorrect answers. However, it is common for a single image to be associated with multiple questions, and LVLMs may still answer other questions correctly even for an adversarial image attacked by a specific question. To address this, we introduce the query-agnostic visual attack (QAVA), which aims to create robust adversarial examples that generate incorrect responses to unspecified and unknown questions. Com"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2504.11038","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/2504.11038/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":"2504.11038","created_at":"2026-07-05T10:49:28.149367+00:00"},{"alias_kind":"arxiv_version","alias_value":"2504.11038v1","created_at":"2026-07-05T10:49:28.149367+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2504.11038","created_at":"2026-07-05T10:49:28.149367+00:00"},{"alias_kind":"pith_short_12","alias_value":"22D77WXEKTPA","created_at":"2026-07-05T10:49:28.149367+00:00"},{"alias_kind":"pith_short_16","alias_value":"22D77WXEKTPAJDZT","created_at":"2026-07-05T10:49:28.149367+00:00"},{"alias_kind":"pith_short_8","alias_value":"22D77WXE","created_at":"2026-07-05T10:49:28.149367+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2507.08982","citing_title":"VIP: Visual Information Protection through Adversarial Attacks on Vision-Language Models","ref_index":41,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/22D77WXEKTPAJDZTVLJPNVQPYN","json":"https://pith.science/pith/22D77WXEKTPAJDZTVLJPNVQPYN.json","graph_json":"https://pith.science/api/pith-number/22D77WXEKTPAJDZTVLJPNVQPYN/graph.json","events_json":"https://pith.science/api/pith-number/22D77WXEKTPAJDZTVLJPNVQPYN/events.json","paper":"https://pith.science/paper/22D77WXE"},"agent_actions":{"view_html":"https://pith.science/pith/22D77WXEKTPAJDZTVLJPNVQPYN","download_json":"https://pith.science/pith/22D77WXEKTPAJDZTVLJPNVQPYN.json","view_paper":"https://pith.science/paper/22D77WXE","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2504.11038&json=true","fetch_graph":"https://pith.science/api/pith-number/22D77WXEKTPAJDZTVLJPNVQPYN/graph.json","fetch_events":"https://pith.science/api/pith-number/22D77WXEKTPAJDZTVLJPNVQPYN/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/22D77WXEKTPAJDZTVLJPNVQPYN/action/timestamp_anchor","attest_storage":"https://pith.science/pith/22D77WXEKTPAJDZTVLJPNVQPYN/action/storage_attestation","attest_author":"https://pith.science/pith/22D77WXEKTPAJDZTVLJPNVQPYN/action/author_attestation","sign_citation":"https://pith.science/pith/22D77WXEKTPAJDZTVLJPNVQPYN/action/citation_signature","submit_replication":"https://pith.science/pith/22D77WXEKTPAJDZTVLJPNVQPYN/action/replication_record"}},"created_at":"2026-07-05T10:49:28.149367+00:00","updated_at":"2026-07-05T10:49:28.149367+00:00"}