{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:BDTU7AU7S3W6VWZM64HUGR3DXS","short_pith_number":"pith:BDTU7AU7","schema_version":"1.0","canonical_sha256":"08e74f829f96edeadb2cf70f434763bc9e1fec32e67d222343272daf00ea07f4","source":{"kind":"arxiv","id":"2506.05394","version":3},"attestation_state":"computed","paper":{"title":"Attacking Attention of Foundation Models Disrupts Downstream Tasks","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CR","authors_text":"Federico Becattini, Hondamunige Prasanna Silva, Lorenzo Seidenari","submitted_at":"2025-06-03T19:42:48Z","abstract_excerpt":"Foundation models represent the most prominent and recent paradigm shift in artificial intelligence. Foundation models are large models, trained on broad data that deliver high accuracy in many downstream tasks, often without fine-tuning. For this reason, models such as CLIP , DINO or Vision Transfomers (ViT), are becoming the bedrock of many industrial AI-powered applications. However, the reliance on pre-trained foundation models also introduces significant security concerns, as these models are vulnerable to adversarial attacks. Such attacks involve deliberately crafted inputs designed to d"},"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":"2506.05394","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2025-06-03T19:42:48Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"89b41f7fa5eaccd056bb3332182d29f77df970000b68f6271834635e64cb292e","abstract_canon_sha256":"f92c849889ea22667113be29bf9c379319c7d3a760e86a732089085b1ce29bbf"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T12:10:58.318350Z","signature_b64":"PVIifsfe3D79QYMIS6/3lRqPa+ggeVUH5xN6vfkrgOZ0c9G/2XGRdHaKr+Nu5C52QYfk9HAx2Rp2voPNmKPABQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"08e74f829f96edeadb2cf70f434763bc9e1fec32e67d222343272daf00ea07f4","last_reissued_at":"2026-07-05T12:10:58.317845Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T12:10:58.317845Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Attacking Attention of Foundation Models Disrupts Downstream Tasks","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CR","authors_text":"Federico Becattini, Hondamunige Prasanna Silva, Lorenzo Seidenari","submitted_at":"2025-06-03T19:42:48Z","abstract_excerpt":"Foundation models represent the most prominent and recent paradigm shift in artificial intelligence. Foundation models are large models, trained on broad data that deliver high accuracy in many downstream tasks, often without fine-tuning. For this reason, models such as CLIP , DINO or Vision Transfomers (ViT), are becoming the bedrock of many industrial AI-powered applications. However, the reliance on pre-trained foundation models also introduces significant security concerns, as these models are vulnerable to adversarial attacks. Such attacks involve deliberately crafted inputs designed to d"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.05394","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/2506.05394/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":"2506.05394","created_at":"2026-07-05T12:10:58.317901+00:00"},{"alias_kind":"arxiv_version","alias_value":"2506.05394v3","created_at":"2026-07-05T12:10:58.317901+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.05394","created_at":"2026-07-05T12:10:58.317901+00:00"},{"alias_kind":"pith_short_12","alias_value":"BDTU7AU7S3W6","created_at":"2026-07-05T12:10:58.317901+00:00"},{"alias_kind":"pith_short_16","alias_value":"BDTU7AU7S3W6VWZM","created_at":"2026-07-05T12:10:58.317901+00:00"},{"alias_kind":"pith_short_8","alias_value":"BDTU7AU7","created_at":"2026-07-05T12:10:58.317901+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/BDTU7AU7S3W6VWZM64HUGR3DXS","json":"https://pith.science/pith/BDTU7AU7S3W6VWZM64HUGR3DXS.json","graph_json":"https://pith.science/api/pith-number/BDTU7AU7S3W6VWZM64HUGR3DXS/graph.json","events_json":"https://pith.science/api/pith-number/BDTU7AU7S3W6VWZM64HUGR3DXS/events.json","paper":"https://pith.science/paper/BDTU7AU7"},"agent_actions":{"view_html":"https://pith.science/pith/BDTU7AU7S3W6VWZM64HUGR3DXS","download_json":"https://pith.science/pith/BDTU7AU7S3W6VWZM64HUGR3DXS.json","view_paper":"https://pith.science/paper/BDTU7AU7","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2506.05394&json=true","fetch_graph":"https://pith.science/api/pith-number/BDTU7AU7S3W6VWZM64HUGR3DXS/graph.json","fetch_events":"https://pith.science/api/pith-number/BDTU7AU7S3W6VWZM64HUGR3DXS/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/BDTU7AU7S3W6VWZM64HUGR3DXS/action/timestamp_anchor","attest_storage":"https://pith.science/pith/BDTU7AU7S3W6VWZM64HUGR3DXS/action/storage_attestation","attest_author":"https://pith.science/pith/BDTU7AU7S3W6VWZM64HUGR3DXS/action/author_attestation","sign_citation":"https://pith.science/pith/BDTU7AU7S3W6VWZM64HUGR3DXS/action/citation_signature","submit_replication":"https://pith.science/pith/BDTU7AU7S3W6VWZM64HUGR3DXS/action/replication_record"}},"created_at":"2026-07-05T12:10:58.317901+00:00","updated_at":"2026-07-05T12:10:58.317901+00:00"}