{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:4NTJONG2IZPNP5XCLPXH3I2AMB","short_pith_number":"pith:4NTJONG2","schema_version":"1.0","canonical_sha256":"e3669734da465ed7f6e25bee7da340605f0ab8fabe51322333829823f2855cba","source":{"kind":"arxiv","id":"2608.02137","version":1},"attestation_state":"computed","paper":{"title":"Two Sides of the Same Coin: Co-Evolving Search for Cross-Task Attacks on Vision-Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Junhao Dong, Mingrong Gong, Xinghua Qu, Xuanhui Lin, Yew-Soon Ong, Yucheng Chen","submitted_at":"2026-08-03T12:24:50Z","abstract_excerpt":"Vision-language models (VLMs) exhibit strong generalization across multimodal tasks but remain vulnerable to adversarial perturbations. Existing attacks typically follow single-trajectory gradient optimization or task-specific objectives, limiting search-space exploration and cross-task transferability. We propose an evolutionary-computation-guided cross-modal attack framework for unified VLMs. The framework adaptively searches both textual and visual spaces. On the textual side, it evolves hard negative semantic embeddings around the source-category representation to provide diverse cross-mod"},"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":"2608.02137","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2026-08-03T12:24:50Z","cross_cats_sorted":[],"title_canon_sha256":"b1d0b4719a20bfd454b6a1dfe3c55a84ff65a28f689f5a580be9c1d3a494aa8e","abstract_canon_sha256":"7b60ce588acc5aac588bff45173842f57c6502a573aa560319ddacdf26458e3f"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-08-04T02:11:05.199115Z","signature_b64":"k1s3FajN45Qcoumymff2HLo/hv9bNGwzj7+t1EGve7OR9dbdr95hp7mQ6NaHse9pLlfTZ/4QKAo0YsM7XFSBDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e3669734da465ed7f6e25bee7da340605f0ab8fabe51322333829823f2855cba","last_reissued_at":"2026-08-04T02:11:05.197615Z","signature_status":"signed_v1","first_computed_at":"2026-08-04T02:11:05.197615Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Two Sides of the Same Coin: Co-Evolving Search for Cross-Task Attacks on Vision-Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Junhao Dong, Mingrong Gong, Xinghua Qu, Xuanhui Lin, Yew-Soon Ong, Yucheng Chen","submitted_at":"2026-08-03T12:24:50Z","abstract_excerpt":"Vision-language models (VLMs) exhibit strong generalization across multimodal tasks but remain vulnerable to adversarial perturbations. Existing attacks typically follow single-trajectory gradient optimization or task-specific objectives, limiting search-space exploration and cross-task transferability. We propose an evolutionary-computation-guided cross-modal attack framework for unified VLMs. The framework adaptively searches both textual and visual spaces. On the textual side, it evolves hard negative semantic embeddings around the source-category representation to provide diverse cross-mod"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2608.02137","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/2608.02137/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":"2608.02137","created_at":"2026-08-04T02:11:05.199125+00:00"},{"alias_kind":"arxiv_version","alias_value":"2608.02137v1","created_at":"2026-08-04T02:11:05.199125+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2608.02137","created_at":"2026-08-04T02:11:05.199125+00:00"},{"alias_kind":"pith_short_12","alias_value":"4NTJONG2IZPN","created_at":"2026-08-04T02:11:05.199125+00:00"},{"alias_kind":"pith_short_16","alias_value":"4NTJONG2IZPNP5XC","created_at":"2026-08-04T02:11:05.199125+00:00"},{"alias_kind":"pith_short_8","alias_value":"4NTJONG2","created_at":"2026-08-04T02:11:05.199125+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/4NTJONG2IZPNP5XCLPXH3I2AMB","json":"https://pith.science/pith/4NTJONG2IZPNP5XCLPXH3I2AMB.json","graph_json":"https://pith.science/api/pith-number/4NTJONG2IZPNP5XCLPXH3I2AMB/graph.json","events_json":"https://pith.science/api/pith-number/4NTJONG2IZPNP5XCLPXH3I2AMB/events.json","paper":"https://pith.science/paper/4NTJONG2"},"agent_actions":{"view_html":"https://pith.science/pith/4NTJONG2IZPNP5XCLPXH3I2AMB","download_json":"https://pith.science/pith/4NTJONG2IZPNP5XCLPXH3I2AMB.json","view_paper":"https://pith.science/paper/4NTJONG2","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2608.02137&json=true","fetch_graph":"https://pith.science/api/pith-number/4NTJONG2IZPNP5XCLPXH3I2AMB/graph.json","fetch_events":"https://pith.science/api/pith-number/4NTJONG2IZPNP5XCLPXH3I2AMB/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/4NTJONG2IZPNP5XCLPXH3I2AMB/action/timestamp_anchor","attest_storage":"https://pith.science/pith/4NTJONG2IZPNP5XCLPXH3I2AMB/action/storage_attestation","attest_author":"https://pith.science/pith/4NTJONG2IZPNP5XCLPXH3I2AMB/action/author_attestation","sign_citation":"https://pith.science/pith/4NTJONG2IZPNP5XCLPXH3I2AMB/action/citation_signature","submit_replication":"https://pith.science/pith/4NTJONG2IZPNP5XCLPXH3I2AMB/action/replication_record"}},"created_at":"2026-08-04T02:11:05.199125+00:00","updated_at":"2026-08-04T02:11:05.199125+00:00"}