{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:MSUT2FXCDCWIFEHHNAPO57IJAL","short_pith_number":"pith:MSUT2FXC","schema_version":"1.0","canonical_sha256":"64a93d16e218ac8290e7681eeefd0902ce38bc34c55e3bcfb5651e9d23464fa7","source":{"kind":"arxiv","id":"2302.09042","version":2},"attestation_state":"computed","paper":{"title":"Privately Customizing Prefinetuning to Better Match User Data in Federated Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.DC"],"primary_cat":"cs.LG","authors_text":"Akshat Shrivastava, Aleksandr Livshits, Charlie Hou, Daniel Lazar, Giulia Fanti, Hongyuan Zhan, Sid Wang","submitted_at":"2023-02-17T18:18:22Z","abstract_excerpt":"In Federated Learning (FL), accessing private client data incurs communication and privacy costs. As a result, FL deployments commonly prefinetune pretrained foundation models on a (large, possibly public) dataset that is held by the central server; they then FL-finetune the model on a private, federated dataset held by clients. Evaluating prefinetuning dataset quality reliably and privately is therefore of high importance. To this end, we propose FreD (Federated Private Fr\\'echet Distance) -- a privately computed distance between a prefinetuning dataset and federated datasets. Intuitively, it"},"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":"2302.09042","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2023-02-17T18:18:22Z","cross_cats_sorted":["cs.AI","cs.DC"],"title_canon_sha256":"799dcadf4715312cd17b48d16e8069d1b7539f60f54ec1188e6ad72f096d97f7","abstract_canon_sha256":"b3da6b15260fc93e4e83cfebdd5f88421ee3526eb0ee45ed652d2a30d7d53a17"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T05:44:53.256231Z","signature_b64":"nNt+ZhRIbm/lJ5szJ/h+Nn6bQIW07EqrpszW5lKv4+swQi0w49VikgZEYYuD4doRx5aFM88vcEfKwQbkvezMCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"64a93d16e218ac8290e7681eeefd0902ce38bc34c55e3bcfb5651e9d23464fa7","last_reissued_at":"2026-07-05T05:44:53.255808Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T05:44:53.255808Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Privately Customizing Prefinetuning to Better Match User Data in Federated Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.DC"],"primary_cat":"cs.LG","authors_text":"Akshat Shrivastava, Aleksandr Livshits, Charlie Hou, Daniel Lazar, Giulia Fanti, Hongyuan Zhan, Sid Wang","submitted_at":"2023-02-17T18:18:22Z","abstract_excerpt":"In Federated Learning (FL), accessing private client data incurs communication and privacy costs. As a result, FL deployments commonly prefinetune pretrained foundation models on a (large, possibly public) dataset that is held by the central server; they then FL-finetune the model on a private, federated dataset held by clients. Evaluating prefinetuning dataset quality reliably and privately is therefore of high importance. To this end, we propose FreD (Federated Private Fr\\'echet Distance) -- a privately computed distance between a prefinetuning dataset and federated datasets. Intuitively, it"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2302.09042","kind":"arxiv","version":2},"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/2302.09042/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":"2302.09042","created_at":"2026-07-05T05:44:53.255867+00:00"},{"alias_kind":"arxiv_version","alias_value":"2302.09042v2","created_at":"2026-07-05T05:44:53.255867+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2302.09042","created_at":"2026-07-05T05:44:53.255867+00:00"},{"alias_kind":"pith_short_12","alias_value":"MSUT2FXCDCWI","created_at":"2026-07-05T05:44:53.255867+00:00"},{"alias_kind":"pith_short_16","alias_value":"MSUT2FXCDCWIFEHH","created_at":"2026-07-05T05:44:53.255867+00:00"},{"alias_kind":"pith_short_8","alias_value":"MSUT2FXC","created_at":"2026-07-05T05:44:53.255867+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2411.16478","citing_title":"Distributed, communication-efficient, and differentially private estimation of KL divergence","ref_index":13,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/MSUT2FXCDCWIFEHHNAPO57IJAL","json":"https://pith.science/pith/MSUT2FXCDCWIFEHHNAPO57IJAL.json","graph_json":"https://pith.science/api/pith-number/MSUT2FXCDCWIFEHHNAPO57IJAL/graph.json","events_json":"https://pith.science/api/pith-number/MSUT2FXCDCWIFEHHNAPO57IJAL/events.json","paper":"https://pith.science/paper/MSUT2FXC"},"agent_actions":{"view_html":"https://pith.science/pith/MSUT2FXCDCWIFEHHNAPO57IJAL","download_json":"https://pith.science/pith/MSUT2FXCDCWIFEHHNAPO57IJAL.json","view_paper":"https://pith.science/paper/MSUT2FXC","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2302.09042&json=true","fetch_graph":"https://pith.science/api/pith-number/MSUT2FXCDCWIFEHHNAPO57IJAL/graph.json","fetch_events":"https://pith.science/api/pith-number/MSUT2FXCDCWIFEHHNAPO57IJAL/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/MSUT2FXCDCWIFEHHNAPO57IJAL/action/timestamp_anchor","attest_storage":"https://pith.science/pith/MSUT2FXCDCWIFEHHNAPO57IJAL/action/storage_attestation","attest_author":"https://pith.science/pith/MSUT2FXCDCWIFEHHNAPO57IJAL/action/author_attestation","sign_citation":"https://pith.science/pith/MSUT2FXCDCWIFEHHNAPO57IJAL/action/citation_signature","submit_replication":"https://pith.science/pith/MSUT2FXCDCWIFEHHNAPO57IJAL/action/replication_record"}},"created_at":"2026-07-05T05:44:53.255867+00:00","updated_at":"2026-07-05T05:44:53.255867+00:00"}