{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:YGGHA4W7M2L5S7HZTHJ2PZXCDO","short_pith_number":"pith:YGGHA4W7","schema_version":"1.0","canonical_sha256":"c18c7072df6697d97cf999d3a7e6e21b803401274f5bf0a5c30b2b0e351d559b","source":{"kind":"arxiv","id":"2509.06992","version":1},"attestation_state":"computed","paper":{"title":"FedAPT: Federated Adversarial Prompt Tuning for Vision-Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Kun Zhai, Siheng Chen, Xingjun Ma, Yu-Gang Jiang","submitted_at":"2025-09-03T03:46:35Z","abstract_excerpt":"Federated Prompt Tuning (FPT) is an efficient method for cross-client collaborative fine-tuning of large Vision-Language Models (VLMs). However, models tuned using FPT are vulnerable to adversarial attacks, leading to misclassification in downstream tasks. In this work, we introduce Federated Adversarial Prompt Tuning (\\textbf{FedAPT}), a novel method designed to enhance the adversarial robustness of FPT. We identify a key issue in FedAPT under non-independent and identically distributed (non-IID) settings: a \\textit{class information gap} between clients and the global model. Clients rely sol"},"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":"2509.06992","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2025-09-03T03:46:35Z","cross_cats_sorted":[],"title_canon_sha256":"9da748e6aec13f23ed011cc3bd568cc233dd60b705a1beb8aa286d88d40ea6d1","abstract_canon_sha256":"32e86ed605def71d0b5e6fde46f80af7b0ff0ad976cec6284a066f5bd2470904"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T12:07:13.637246Z","signature_b64":"D3OofGc1+ZFM3YQJSmSbvIqmxPpblpA2R5DZGpm9FKydRG2Trf9glRuQv+boKHenNKjcP+3mUHD8k7jjxGITCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c18c7072df6697d97cf999d3a7e6e21b803401274f5bf0a5c30b2b0e351d559b","last_reissued_at":"2026-07-05T12:07:13.636755Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T12:07:13.636755Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"FedAPT: Federated Adversarial Prompt Tuning for Vision-Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Kun Zhai, Siheng Chen, Xingjun Ma, Yu-Gang Jiang","submitted_at":"2025-09-03T03:46:35Z","abstract_excerpt":"Federated Prompt Tuning (FPT) is an efficient method for cross-client collaborative fine-tuning of large Vision-Language Models (VLMs). However, models tuned using FPT are vulnerable to adversarial attacks, leading to misclassification in downstream tasks. In this work, we introduce Federated Adversarial Prompt Tuning (\\textbf{FedAPT}), a novel method designed to enhance the adversarial robustness of FPT. We identify a key issue in FedAPT under non-independent and identically distributed (non-IID) settings: a \\textit{class information gap} between clients and the global model. Clients rely sol"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2509.06992","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/2509.06992/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":"2509.06992","created_at":"2026-07-05T12:07:13.636814+00:00"},{"alias_kind":"arxiv_version","alias_value":"2509.06992v1","created_at":"2026-07-05T12:07:13.636814+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2509.06992","created_at":"2026-07-05T12:07:13.636814+00:00"},{"alias_kind":"pith_short_12","alias_value":"YGGHA4W7M2L5","created_at":"2026-07-05T12:07:13.636814+00:00"},{"alias_kind":"pith_short_16","alias_value":"YGGHA4W7M2L5S7HZ","created_at":"2026-07-05T12:07:13.636814+00:00"},{"alias_kind":"pith_short_8","alias_value":"YGGHA4W7","created_at":"2026-07-05T12:07:13.636814+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/YGGHA4W7M2L5S7HZTHJ2PZXCDO","json":"https://pith.science/pith/YGGHA4W7M2L5S7HZTHJ2PZXCDO.json","graph_json":"https://pith.science/api/pith-number/YGGHA4W7M2L5S7HZTHJ2PZXCDO/graph.json","events_json":"https://pith.science/api/pith-number/YGGHA4W7M2L5S7HZTHJ2PZXCDO/events.json","paper":"https://pith.science/paper/YGGHA4W7"},"agent_actions":{"view_html":"https://pith.science/pith/YGGHA4W7M2L5S7HZTHJ2PZXCDO","download_json":"https://pith.science/pith/YGGHA4W7M2L5S7HZTHJ2PZXCDO.json","view_paper":"https://pith.science/paper/YGGHA4W7","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2509.06992&json=true","fetch_graph":"https://pith.science/api/pith-number/YGGHA4W7M2L5S7HZTHJ2PZXCDO/graph.json","fetch_events":"https://pith.science/api/pith-number/YGGHA4W7M2L5S7HZTHJ2PZXCDO/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/YGGHA4W7M2L5S7HZTHJ2PZXCDO/action/timestamp_anchor","attest_storage":"https://pith.science/pith/YGGHA4W7M2L5S7HZTHJ2PZXCDO/action/storage_attestation","attest_author":"https://pith.science/pith/YGGHA4W7M2L5S7HZTHJ2PZXCDO/action/author_attestation","sign_citation":"https://pith.science/pith/YGGHA4W7M2L5S7HZTHJ2PZXCDO/action/citation_signature","submit_replication":"https://pith.science/pith/YGGHA4W7M2L5S7HZTHJ2PZXCDO/action/replication_record"}},"created_at":"2026-07-05T12:07:13.636814+00:00","updated_at":"2026-07-05T12:07:13.636814+00:00"}