{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:ID4SF467P7F4DESWBEA4O3BM4N","short_pith_number":"pith:ID4SF467","schema_version":"1.0","canonical_sha256":"40f922f3df7fcbc192560901c76c2ce343ff39ec586920b2d997ef58997a31dc","source":{"kind":"arxiv","id":"2608.03231","version":1},"attestation_state":"computed","paper":{"title":"Structure-Aware Robust Fine-Tuning: Defending Vision-Language-Action Robots Against Physical Attention Hijacking","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.RO","authors_text":"Dongfu Yin, F. Richard Yu, Jinquan Zhang, Run Yang, Yufeng Yan, Zhen Tian","submitted_at":"2026-08-04T07:03:05Z","abstract_excerpt":"Vision-Language-Action (VLA) policies promise general robotic manipulation, but their robustness against physical-world attacks remains fragile. In particular, we show that physically realizable adversarial patches can reliably induce failures by triggering a mechanism we call policy-critical action-to-vision attention hijacking, where action-conditioned attention is diverted from task-relevant regions to a localized patch. To demonstrate the threat, we propose Attention-Guided Semantic Disruption (AGSD), an Expectation-over-Transformation (EOT) optimized printable patch that jointly (i) conce"},"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":"2608.03231","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.RO","submitted_at":"2026-08-04T07:03:05Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"4adc8c04069a24f003c37ec8dd33f61e7459faa6a64e587a057d2fb032e1f929","abstract_canon_sha256":"ccc39b5c450de3af3488697870d4d9881d0159f6fd0844819f855343a211b858"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-08-05T00:46:20.742194Z","signature_b64":"2jQUIpS5dBRicZBlzO/TnELk32lbgA3dSR9HKQ+d8C2bEbPioNcg+mpgE+T6rq3yWieVwuQ4/TZqlKX5kyQaCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"40f922f3df7fcbc192560901c76c2ce343ff39ec586920b2d997ef58997a31dc","last_reissued_at":"2026-08-05T00:46:20.739032Z","signature_status":"signed_v1","first_computed_at":"2026-08-05T00:46:20.739032Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Structure-Aware Robust Fine-Tuning: Defending Vision-Language-Action Robots Against Physical Attention Hijacking","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.RO","authors_text":"Dongfu Yin, F. Richard Yu, Jinquan Zhang, Run Yang, Yufeng Yan, Zhen Tian","submitted_at":"2026-08-04T07:03:05Z","abstract_excerpt":"Vision-Language-Action (VLA) policies promise general robotic manipulation, but their robustness against physical-world attacks remains fragile. In particular, we show that physically realizable adversarial patches can reliably induce failures by triggering a mechanism we call policy-critical action-to-vision attention hijacking, where action-conditioned attention is diverted from task-relevant regions to a localized patch. To demonstrate the threat, we propose Attention-Guided Semantic Disruption (AGSD), an Expectation-over-Transformation (EOT) optimized printable patch that jointly (i) conce"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2608.03231","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/2608.03231/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":"2608.03231","created_at":"2026-08-05T00:46:20.739980+00:00"},{"alias_kind":"arxiv_version","alias_value":"2608.03231v1","created_at":"2026-08-05T00:46:20.739980+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2608.03231","created_at":"2026-08-05T00:46:20.739980+00:00"},{"alias_kind":"pith_short_12","alias_value":"ID4SF467P7F4","created_at":"2026-08-05T00:46:20.739980+00:00"},{"alias_kind":"pith_short_16","alias_value":"ID4SF467P7F4DESW","created_at":"2026-08-05T00:46:20.739980+00:00"},{"alias_kind":"pith_short_8","alias_value":"ID4SF467","created_at":"2026-08-05T00:46:20.739980+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/ID4SF467P7F4DESWBEA4O3BM4N","json":"https://pith.science/pith/ID4SF467P7F4DESWBEA4O3BM4N.json","graph_json":"https://pith.science/api/pith-number/ID4SF467P7F4DESWBEA4O3BM4N/graph.json","events_json":"https://pith.science/api/pith-number/ID4SF467P7F4DESWBEA4O3BM4N/events.json","paper":"https://pith.science/paper/ID4SF467"},"agent_actions":{"view_html":"https://pith.science/pith/ID4SF467P7F4DESWBEA4O3BM4N","download_json":"https://pith.science/pith/ID4SF467P7F4DESWBEA4O3BM4N.json","view_paper":"https://pith.science/paper/ID4SF467","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2608.03231&json=true","fetch_graph":"https://pith.science/api/pith-number/ID4SF467P7F4DESWBEA4O3BM4N/graph.json","fetch_events":"https://pith.science/api/pith-number/ID4SF467P7F4DESWBEA4O3BM4N/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/ID4SF467P7F4DESWBEA4O3BM4N/action/timestamp_anchor","attest_storage":"https://pith.science/pith/ID4SF467P7F4DESWBEA4O3BM4N/action/storage_attestation","attest_author":"https://pith.science/pith/ID4SF467P7F4DESWBEA4O3BM4N/action/author_attestation","sign_citation":"https://pith.science/pith/ID4SF467P7F4DESWBEA4O3BM4N/action/citation_signature","submit_replication":"https://pith.science/pith/ID4SF467P7F4DESWBEA4O3BM4N/action/replication_record"}},"created_at":"2026-08-05T00:46:20.739980+00:00","updated_at":"2026-08-05T00:46:20.739980+00:00"}