{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:DH6KRNHCYNHPMEBSYLWXMI3LDC","short_pith_number":"pith:DH6KRNHC","schema_version":"1.0","canonical_sha256":"19fca8b4e2c34ef61032c2ed76236b18809c44f1a66e0e4e39be141a6258e78f","source":{"kind":"arxiv","id":"2410.22888","version":1},"attestation_state":"computed","paper":{"title":"Effective and Efficient Adversarial Detection for Vision-Language Models via A Single Vector","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL","cs.CR"],"primary_cat":"cs.CV","authors_text":"Fengbin Zhu, Jiancheng Lv, Jingkun Tang, Pan Zhou, Tat-Seng Chua, Wenqiang Lei, Youcheng Huang","submitted_at":"2024-10-30T10:33:10Z","abstract_excerpt":"Visual Language Models (VLMs) are vulnerable to adversarial attacks, especially those from adversarial images, which is however under-explored in literature. To facilitate research on this critical safety problem, we first construct a new laRge-scale Adervsarial images dataset with Diverse hArmful Responses (RADAR), given that existing datasets are either small-scale or only contain limited types of harmful responses. With the new RADAR dataset, we further develop a novel and effective iN-time Embedding-based AdveRSarial Image DEtection (NEARSIDE) method, which exploits a single vector that di"},"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":"2410.22888","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2024-10-30T10:33:10Z","cross_cats_sorted":["cs.CL","cs.CR"],"title_canon_sha256":"711770155c806b49fd5cb046227f2eedd6480f9330b0da7daa2d9da452dec79b","abstract_canon_sha256":"96bd779e0952482fe1e261bea5384fefef38803c54e79a4a1455d13a34f59c7b"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:28:37.577014Z","signature_b64":"bdMS2jl0mhR4JVp2/NTyPRbmj5LmCReXK/TXCcgZ3jKULTX002niPJB8yrYg4pLWJPibxZAK8LzeHxG0bqFYBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"19fca8b4e2c34ef61032c2ed76236b18809c44f1a66e0e4e39be141a6258e78f","last_reissued_at":"2026-07-05T09:28:37.576539Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:28:37.576539Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Effective and Efficient Adversarial Detection for Vision-Language Models via A Single Vector","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL","cs.CR"],"primary_cat":"cs.CV","authors_text":"Fengbin Zhu, Jiancheng Lv, Jingkun Tang, Pan Zhou, Tat-Seng Chua, Wenqiang Lei, Youcheng Huang","submitted_at":"2024-10-30T10:33:10Z","abstract_excerpt":"Visual Language Models (VLMs) are vulnerable to adversarial attacks, especially those from adversarial images, which is however under-explored in literature. To facilitate research on this critical safety problem, we first construct a new laRge-scale Adervsarial images dataset with Diverse hArmful Responses (RADAR), given that existing datasets are either small-scale or only contain limited types of harmful responses. With the new RADAR dataset, we further develop a novel and effective iN-time Embedding-based AdveRSarial Image DEtection (NEARSIDE) method, which exploits a single vector that di"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.22888","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/2410.22888/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":"2410.22888","created_at":"2026-07-05T09:28:37.576597+00:00"},{"alias_kind":"arxiv_version","alias_value":"2410.22888v1","created_at":"2026-07-05T09:28:37.576597+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.22888","created_at":"2026-07-05T09:28:37.576597+00:00"},{"alias_kind":"pith_short_12","alias_value":"DH6KRNHCYNHP","created_at":"2026-07-05T09:28:37.576597+00:00"},{"alias_kind":"pith_short_16","alias_value":"DH6KRNHCYNHPMEBS","created_at":"2026-07-05T09:28:37.576597+00:00"},{"alias_kind":"pith_short_8","alias_value":"DH6KRNHC","created_at":"2026-07-05T09:28:37.576597+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2507.00052","citing_title":"VSF-Med:A Vulnerability Scoring Framework for Medical Vision-Language Models","ref_index":7,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/DH6KRNHCYNHPMEBSYLWXMI3LDC","json":"https://pith.science/pith/DH6KRNHCYNHPMEBSYLWXMI3LDC.json","graph_json":"https://pith.science/api/pith-number/DH6KRNHCYNHPMEBSYLWXMI3LDC/graph.json","events_json":"https://pith.science/api/pith-number/DH6KRNHCYNHPMEBSYLWXMI3LDC/events.json","paper":"https://pith.science/paper/DH6KRNHC"},"agent_actions":{"view_html":"https://pith.science/pith/DH6KRNHCYNHPMEBSYLWXMI3LDC","download_json":"https://pith.science/pith/DH6KRNHCYNHPMEBSYLWXMI3LDC.json","view_paper":"https://pith.science/paper/DH6KRNHC","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2410.22888&json=true","fetch_graph":"https://pith.science/api/pith-number/DH6KRNHCYNHPMEBSYLWXMI3LDC/graph.json","fetch_events":"https://pith.science/api/pith-number/DH6KRNHCYNHPMEBSYLWXMI3LDC/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/DH6KRNHCYNHPMEBSYLWXMI3LDC/action/timestamp_anchor","attest_storage":"https://pith.science/pith/DH6KRNHCYNHPMEBSYLWXMI3LDC/action/storage_attestation","attest_author":"https://pith.science/pith/DH6KRNHCYNHPMEBSYLWXMI3LDC/action/author_attestation","sign_citation":"https://pith.science/pith/DH6KRNHCYNHPMEBSYLWXMI3LDC/action/citation_signature","submit_replication":"https://pith.science/pith/DH6KRNHCYNHPMEBSYLWXMI3LDC/action/replication_record"}},"created_at":"2026-07-05T09:28:37.576597+00:00","updated_at":"2026-07-05T09:28:37.576597+00:00"}