{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:PMQSKPWHJ2VTT3QOKHLMSS7UIZ","short_pith_number":"pith:PMQSKPWH","schema_version":"1.0","canonical_sha256":"7b21253ec74eab39ee0e51d6c94bf4467347d30d52fec3252ad3eed1c7a87639","source":{"kind":"arxiv","id":"2410.05346","version":3},"attestation_state":"computed","paper":{"title":"AnyAttack: Towards Large-scale Self-supervised Adversarial Attacks on Vision-language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Dit-Yan Yeung, Jiaming Zhang, Jitao Sang, Junhong Ye, Xingjun Ma, Yige Li, Yunfan Yang, Yunhao Chen","submitted_at":"2024-10-07T09:45:18Z","abstract_excerpt":"Due to their multimodal capabilities, Vision-Language Models (VLMs) have found numerous impactful applications in real-world scenarios. However, recent studies have revealed that VLMs are vulnerable to image-based adversarial attacks. Traditional targeted adversarial attacks require specific targets and labels, limiting their real-world impact.We present AnyAttack, a self-supervised framework that transcends the limitations of conventional attacks through a novel foundation model approach. By pre-training on the massive LAION-400M dataset without label supervision, AnyAttack achieves unprecede"},"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":"2410.05346","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-10-07T09:45:18Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"7a1ff5996ce95273c60866184ca393b476a4861a156ec1c8b547f75ea0d35394","abstract_canon_sha256":"c304fcf61600618fefbaeb5f0a7075b6fe0ef37f42603558f5735d6fad52679e"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:40:36.127367Z","signature_b64":"+W0imLuKRfox2MQ/hZTPPHLhtDpqN0u0urd6XCDbvNRRh8I640C/jL2a8xHc0DizOdA+GEnSvnvyqhP8vv0zBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"7b21253ec74eab39ee0e51d6c94bf4467347d30d52fec3252ad3eed1c7a87639","last_reissued_at":"2026-07-05T10:40:36.126856Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:40:36.126856Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"AnyAttack: Towards Large-scale Self-supervised Adversarial Attacks on Vision-language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Dit-Yan Yeung, Jiaming Zhang, Jitao Sang, Junhong Ye, Xingjun Ma, Yige Li, Yunfan Yang, Yunhao Chen","submitted_at":"2024-10-07T09:45:18Z","abstract_excerpt":"Due to their multimodal capabilities, Vision-Language Models (VLMs) have found numerous impactful applications in real-world scenarios. However, recent studies have revealed that VLMs are vulnerable to image-based adversarial attacks. Traditional targeted adversarial attacks require specific targets and labels, limiting their real-world impact.We present AnyAttack, a self-supervised framework that transcends the limitations of conventional attacks through a novel foundation model approach. By pre-training on the massive LAION-400M dataset without label supervision, AnyAttack achieves unprecede"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.05346","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/2410.05346/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":"2410.05346","created_at":"2026-07-05T10:40:36.126915+00:00"},{"alias_kind":"arxiv_version","alias_value":"2410.05346v3","created_at":"2026-07-05T10:40:36.126915+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.05346","created_at":"2026-07-05T10:40:36.126915+00:00"},{"alias_kind":"pith_short_12","alias_value":"PMQSKPWHJ2VT","created_at":"2026-07-05T10:40:36.126915+00:00"},{"alias_kind":"pith_short_16","alias_value":"PMQSKPWHJ2VTT3QO","created_at":"2026-07-05T10:40:36.126915+00:00"},{"alias_kind":"pith_short_8","alias_value":"PMQSKPWH","created_at":"2026-07-05T10:40:36.126915+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.26566","citing_title":"Adversarial Diffusion Across Modalities: A Fusion Survey of Attacks, Defenses, and Evaluation for Text, Vision, and Vision-Language Models","ref_index":79,"is_internal_anchor":false},{"citing_arxiv_id":"2411.18275","citing_title":"Visual Adversarial Attack on Vision-Language Models for Autonomous Driving","ref_index":61,"is_internal_anchor":false},{"citing_arxiv_id":"2502.05206","citing_title":"Safety at Scale: A Comprehensive Survey of Large Model and Agent Safety","ref_index":273,"is_internal_anchor":false},{"citing_arxiv_id":"2605.01449","citing_title":"VisInject: Disruption != Injection -- A Dual-Dimension Evaluation of Universal Adversarial Attacks on Vision-Language Models","ref_index":31,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/PMQSKPWHJ2VTT3QOKHLMSS7UIZ","json":"https://pith.science/pith/PMQSKPWHJ2VTT3QOKHLMSS7UIZ.json","graph_json":"https://pith.science/api/pith-number/PMQSKPWHJ2VTT3QOKHLMSS7UIZ/graph.json","events_json":"https://pith.science/api/pith-number/PMQSKPWHJ2VTT3QOKHLMSS7UIZ/events.json","paper":"https://pith.science/paper/PMQSKPWH"},"agent_actions":{"view_html":"https://pith.science/pith/PMQSKPWHJ2VTT3QOKHLMSS7UIZ","download_json":"https://pith.science/pith/PMQSKPWHJ2VTT3QOKHLMSS7UIZ.json","view_paper":"https://pith.science/paper/PMQSKPWH","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2410.05346&json=true","fetch_graph":"https://pith.science/api/pith-number/PMQSKPWHJ2VTT3QOKHLMSS7UIZ/graph.json","fetch_events":"https://pith.science/api/pith-number/PMQSKPWHJ2VTT3QOKHLMSS7UIZ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/PMQSKPWHJ2VTT3QOKHLMSS7UIZ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/PMQSKPWHJ2VTT3QOKHLMSS7UIZ/action/storage_attestation","attest_author":"https://pith.science/pith/PMQSKPWHJ2VTT3QOKHLMSS7UIZ/action/author_attestation","sign_citation":"https://pith.science/pith/PMQSKPWHJ2VTT3QOKHLMSS7UIZ/action/citation_signature","submit_replication":"https://pith.science/pith/PMQSKPWHJ2VTT3QOKHLMSS7UIZ/action/replication_record"}},"created_at":"2026-07-05T10:40:36.126915+00:00","updated_at":"2026-07-05T10:40:36.126915+00:00"}