{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:QGKYJY5DYEKTHJK64O3ITYZZZ6","short_pith_number":"pith:QGKYJY5D","schema_version":"1.0","canonical_sha256":"819584e3a3c11533a55ee3b689e339cf81694c6fe8dbfedb96fc06d725e94dc8","source":{"kind":"arxiv","id":"2508.20452","version":1},"attestation_state":"computed","paper":{"title":"Evaluating Differentially Private Generation of Domain-Specific Text","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CR"],"primary_cat":"cs.LG","authors_text":"Anil A Bharath, Goran Nenadic, Iqra Zahid, Jie Zhang, Siew-Kei Lam, Srinivasan Nandakumar, Viktor Schlegel, Warren Del-Pinto, Yidan Sun, Yuping Wu","submitted_at":"2025-08-28T05:57:47Z","abstract_excerpt":"Generative AI offers transformative potential for high-stakes domains such as healthcare and finance, yet privacy and regulatory barriers hinder the use of real-world data. To address this, differentially private synthetic data generation has emerged as a promising alternative. In this work, we introduce a unified benchmark to systematically evaluate the utility and fidelity of text datasets generated under formal Differential Privacy (DP) guarantees. Our benchmark addresses key challenges in domain-specific benchmarking, including choice of representative data and realistic privacy budgets, a"},"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":"2508.20452","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-08-28T05:57:47Z","cross_cats_sorted":["cs.AI","cs.CR"],"title_canon_sha256":"7b07db3dca5b48cd142d3b93626a934c27718b392127cc18ac3c0a028ae873d8","abstract_canon_sha256":"bbb37d8121427b8e7c43f3dc21a0f0376f3c0db77779a3124d502955e4dc4dcf"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T12:01:26.820445Z","signature_b64":"/BtIGHDcTc5JJDPIif4noaHFjIAoQ56z2jY9iIzQbqH1K1KBgD89TkWLokqq4p1wHjJfZ3yEDoZiCkDPenkZAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"819584e3a3c11533a55ee3b689e339cf81694c6fe8dbfedb96fc06d725e94dc8","last_reissued_at":"2026-07-05T12:01:26.819897Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T12:01:26.819897Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Evaluating Differentially Private Generation of Domain-Specific Text","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CR"],"primary_cat":"cs.LG","authors_text":"Anil A Bharath, Goran Nenadic, Iqra Zahid, Jie Zhang, Siew-Kei Lam, Srinivasan Nandakumar, Viktor Schlegel, Warren Del-Pinto, Yidan Sun, Yuping Wu","submitted_at":"2025-08-28T05:57:47Z","abstract_excerpt":"Generative AI offers transformative potential for high-stakes domains such as healthcare and finance, yet privacy and regulatory barriers hinder the use of real-world data. To address this, differentially private synthetic data generation has emerged as a promising alternative. In this work, we introduce a unified benchmark to systematically evaluate the utility and fidelity of text datasets generated under formal Differential Privacy (DP) guarantees. Our benchmark addresses key challenges in domain-specific benchmarking, including choice of representative data and realistic privacy budgets, a"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2508.20452","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/2508.20452/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":"2508.20452","created_at":"2026-07-05T12:01:26.819967+00:00"},{"alias_kind":"arxiv_version","alias_value":"2508.20452v1","created_at":"2026-07-05T12:01:26.819967+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2508.20452","created_at":"2026-07-05T12:01:26.819967+00:00"},{"alias_kind":"pith_short_12","alias_value":"QGKYJY5DYEKT","created_at":"2026-07-05T12:01:26.819967+00:00"},{"alias_kind":"pith_short_16","alias_value":"QGKYJY5DYEKTHJK6","created_at":"2026-07-05T12:01:26.819967+00:00"},{"alias_kind":"pith_short_8","alias_value":"QGKYJY5D","created_at":"2026-07-05T12:01:26.819967+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2509.14594","citing_title":"SynBench: A Benchmark for Differentially Private Text Generation","ref_index":40,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/QGKYJY5DYEKTHJK64O3ITYZZZ6","json":"https://pith.science/pith/QGKYJY5DYEKTHJK64O3ITYZZZ6.json","graph_json":"https://pith.science/api/pith-number/QGKYJY5DYEKTHJK64O3ITYZZZ6/graph.json","events_json":"https://pith.science/api/pith-number/QGKYJY5DYEKTHJK64O3ITYZZZ6/events.json","paper":"https://pith.science/paper/QGKYJY5D"},"agent_actions":{"view_html":"https://pith.science/pith/QGKYJY5DYEKTHJK64O3ITYZZZ6","download_json":"https://pith.science/pith/QGKYJY5DYEKTHJK64O3ITYZZZ6.json","view_paper":"https://pith.science/paper/QGKYJY5D","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2508.20452&json=true","fetch_graph":"https://pith.science/api/pith-number/QGKYJY5DYEKTHJK64O3ITYZZZ6/graph.json","fetch_events":"https://pith.science/api/pith-number/QGKYJY5DYEKTHJK64O3ITYZZZ6/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/QGKYJY5DYEKTHJK64O3ITYZZZ6/action/timestamp_anchor","attest_storage":"https://pith.science/pith/QGKYJY5DYEKTHJK64O3ITYZZZ6/action/storage_attestation","attest_author":"https://pith.science/pith/QGKYJY5DYEKTHJK64O3ITYZZZ6/action/author_attestation","sign_citation":"https://pith.science/pith/QGKYJY5DYEKTHJK64O3ITYZZZ6/action/citation_signature","submit_replication":"https://pith.science/pith/QGKYJY5DYEKTHJK64O3ITYZZZ6/action/replication_record"}},"created_at":"2026-07-05T12:01:26.819967+00:00","updated_at":"2026-07-05T12:01:26.819967+00:00"}