{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:WVQ6GD2M3C6XZ33SMVOO6VITDD","short_pith_number":"pith:WVQ6GD2M","schema_version":"1.0","canonical_sha256":"b561e30f4cd8bd7cef72655cef551318cc46d51bcb34500adfc350ed31461109","source":{"kind":"arxiv","id":"2607.21417","version":1},"attestation_state":"computed","paper":{"title":"Towards Privacy-Preserving Federated Prompt Tuning under Data Heterogeneity: A Subspace-Decomposed Expert Approach","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Hainan Zhang, Qinnan Zhang, Xiaodong Li, Yifan Sun, Yihao Guo, Yongxin Tong, Yuhua Wang, Yuxiang Jia, Zhiming Zheng","submitted_at":"2026-07-23T15:22:48Z","abstract_excerpt":"Federated prompt tuning (FPT) enables collaborative adaptation of vision--language models (VLMs) using lightweight prompts. Existing methods often address heterogeneity and privacy through a split-prompt design under local differential privacy (DP), combining a shared prompt for global transfer with private prompts for local adaptation. However, a single shared prompt may over-smooth diverse transferable knowledge, weakening the balance between personalization and generalization. Multi-expert prompts (MEPs) can better capture this diversity, but enlarge the communicated space, increasing DP no"},"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":"2607.21417","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2026-07-23T15:22:48Z","cross_cats_sorted":[],"title_canon_sha256":"ae4474d5d16d42b526c9841d27dc34db0dd9c96c23eea391edd584b71e6f78e7","abstract_canon_sha256":"5340b0ff879edbd3cd4a803edf30315185e87def57a0986b71ae33a7520799c1"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-24T01:24:31.393052Z","signature_b64":"hEcWxk9biW1uCuyRzd0jiury9B8jSwJiJHPNw5/jAldBre2zkmOJIZ7rH+2kolWl/V455LnwY2jSc6yh7gy5BA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b561e30f4cd8bd7cef72655cef551318cc46d51bcb34500adfc350ed31461109","last_reissued_at":"2026-07-24T01:24:31.392142Z","signature_status":"signed_v1","first_computed_at":"2026-07-24T01:24:31.392142Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Towards Privacy-Preserving Federated Prompt Tuning under Data Heterogeneity: A Subspace-Decomposed Expert Approach","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Hainan Zhang, Qinnan Zhang, Xiaodong Li, Yifan Sun, Yihao Guo, Yongxin Tong, Yuhua Wang, Yuxiang Jia, Zhiming Zheng","submitted_at":"2026-07-23T15:22:48Z","abstract_excerpt":"Federated prompt tuning (FPT) enables collaborative adaptation of vision--language models (VLMs) using lightweight prompts. Existing methods often address heterogeneity and privacy through a split-prompt design under local differential privacy (DP), combining a shared prompt for global transfer with private prompts for local adaptation. However, a single shared prompt may over-smooth diverse transferable knowledge, weakening the balance between personalization and generalization. Multi-expert prompts (MEPs) can better capture this diversity, but enlarge the communicated space, increasing DP no"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.21417","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/2607.21417/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":"2607.21417","created_at":"2026-07-24T01:24:31.392608+00:00"},{"alias_kind":"arxiv_version","alias_value":"2607.21417v1","created_at":"2026-07-24T01:24:31.392608+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.21417","created_at":"2026-07-24T01:24:31.392608+00:00"},{"alias_kind":"pith_short_12","alias_value":"WVQ6GD2M3C6X","created_at":"2026-07-24T01:24:31.392608+00:00"},{"alias_kind":"pith_short_16","alias_value":"WVQ6GD2M3C6XZ33S","created_at":"2026-07-24T01:24:31.392608+00:00"},{"alias_kind":"pith_short_8","alias_value":"WVQ6GD2M","created_at":"2026-07-24T01:24:31.392608+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/WVQ6GD2M3C6XZ33SMVOO6VITDD","json":"https://pith.science/pith/WVQ6GD2M3C6XZ33SMVOO6VITDD.json","graph_json":"https://pith.science/api/pith-number/WVQ6GD2M3C6XZ33SMVOO6VITDD/graph.json","events_json":"https://pith.science/api/pith-number/WVQ6GD2M3C6XZ33SMVOO6VITDD/events.json","paper":"https://pith.science/paper/WVQ6GD2M"},"agent_actions":{"view_html":"https://pith.science/pith/WVQ6GD2M3C6XZ33SMVOO6VITDD","download_json":"https://pith.science/pith/WVQ6GD2M3C6XZ33SMVOO6VITDD.json","view_paper":"https://pith.science/paper/WVQ6GD2M","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2607.21417&json=true","fetch_graph":"https://pith.science/api/pith-number/WVQ6GD2M3C6XZ33SMVOO6VITDD/graph.json","fetch_events":"https://pith.science/api/pith-number/WVQ6GD2M3C6XZ33SMVOO6VITDD/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/WVQ6GD2M3C6XZ33SMVOO6VITDD/action/timestamp_anchor","attest_storage":"https://pith.science/pith/WVQ6GD2M3C6XZ33SMVOO6VITDD/action/storage_attestation","attest_author":"https://pith.science/pith/WVQ6GD2M3C6XZ33SMVOO6VITDD/action/author_attestation","sign_citation":"https://pith.science/pith/WVQ6GD2M3C6XZ33SMVOO6VITDD/action/citation_signature","submit_replication":"https://pith.science/pith/WVQ6GD2M3C6XZ33SMVOO6VITDD/action/replication_record"}},"created_at":"2026-07-24T01:24:31.392608+00:00","updated_at":"2026-07-24T01:24:31.392608+00:00"}