{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:RUYIDA7FUGGDF7YKQO3WHHPWAV","short_pith_number":"pith:RUYIDA7F","schema_version":"1.0","canonical_sha256":"8d308183e5a18c32ff0a83b7639df605402fbf435295eb4358d1d18a44651771","source":{"kind":"arxiv","id":"2210.14348","version":3},"attestation_state":"computed","paper":{"title":"Synthetic Text Generation with Differential Privacy: A Simple and Practical Recipe","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CR"],"primary_cat":"cs.CL","authors_text":"David Levitan, Girish Kumar, Hoda Shajari, Huan Sun, Huseyin A. Inan, Julia McAnallen, Robert Sim, Xiang Yue, Xuechen Li","submitted_at":"2022-10-25T21:21:17Z","abstract_excerpt":"Privacy concerns have attracted increasing attention in data-driven products due to the tendency of machine learning models to memorize sensitive training data. Generating synthetic versions of such data with a formal privacy guarantee, such as differential privacy (DP), provides a promising path to mitigating these privacy concerns, but previous approaches in this direction have typically failed to produce synthetic data of high quality. In this work, we show that a simple and practical recipe in the text domain is effective: simply fine-tuning a pretrained generative language model with DP e"},"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":"2210.14348","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2022-10-25T21:21:17Z","cross_cats_sorted":["cs.CR"],"title_canon_sha256":"3b5af1b4785376914edfb866198b0813c8c5e82f76694ff8a1e6b03dd6602927","abstract_canon_sha256":"cf7e658929045593882642abe46c723f108436962171a4ba8275fe2dd964a4a2"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T06:32:02.077100Z","signature_b64":"QZiPQD2NdU74AVbQIKsuMz9e1AxVe+mDp+OuR82QCAlWGv+w7BTEN7e5tggVg4nkjDTdL8+67rbcnELKIcHtDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"8d308183e5a18c32ff0a83b7639df605402fbf435295eb4358d1d18a44651771","last_reissued_at":"2026-07-05T06:32:02.076578Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T06:32:02.076578Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Synthetic Text Generation with Differential Privacy: A Simple and Practical Recipe","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CR"],"primary_cat":"cs.CL","authors_text":"David Levitan, Girish Kumar, Hoda Shajari, Huan Sun, Huseyin A. Inan, Julia McAnallen, Robert Sim, Xiang Yue, Xuechen Li","submitted_at":"2022-10-25T21:21:17Z","abstract_excerpt":"Privacy concerns have attracted increasing attention in data-driven products due to the tendency of machine learning models to memorize sensitive training data. Generating synthetic versions of such data with a formal privacy guarantee, such as differential privacy (DP), provides a promising path to mitigating these privacy concerns, but previous approaches in this direction have typically failed to produce synthetic data of high quality. In this work, we show that a simple and practical recipe in the text domain is effective: simply fine-tuning a pretrained generative language model with DP e"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2210.14348","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/2210.14348/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":"2210.14348","created_at":"2026-07-05T06:32:02.076650+00:00"},{"alias_kind":"arxiv_version","alias_value":"2210.14348v3","created_at":"2026-07-05T06:32:02.076650+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2210.14348","created_at":"2026-07-05T06:32:02.076650+00:00"},{"alias_kind":"pith_short_12","alias_value":"RUYIDA7FUGGD","created_at":"2026-07-05T06:32:02.076650+00:00"},{"alias_kind":"pith_short_16","alias_value":"RUYIDA7FUGGDF7YK","created_at":"2026-07-05T06:32:02.076650+00:00"},{"alias_kind":"pith_short_8","alias_value":"RUYIDA7F","created_at":"2026-07-05T06:32:02.076650+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.00426","citing_title":"Canonicalized Stable-List Replay for Private Federated Continual Learning over Language-Model Embeddings","ref_index":48,"is_internal_anchor":false},{"citing_arxiv_id":"2507.02974","citing_title":"InvisibleInk: High-Utility and Low-Cost Text Generation with Differential Privacy","ref_index":40,"is_internal_anchor":false},{"citing_arxiv_id":"2601.10237","citing_title":"Fundamental Limitations of Favorable Privacy-Utility Guarantees for DP-SGD","ref_index":66,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/RUYIDA7FUGGDF7YKQO3WHHPWAV","json":"https://pith.science/pith/RUYIDA7FUGGDF7YKQO3WHHPWAV.json","graph_json":"https://pith.science/api/pith-number/RUYIDA7FUGGDF7YKQO3WHHPWAV/graph.json","events_json":"https://pith.science/api/pith-number/RUYIDA7FUGGDF7YKQO3WHHPWAV/events.json","paper":"https://pith.science/paper/RUYIDA7F"},"agent_actions":{"view_html":"https://pith.science/pith/RUYIDA7FUGGDF7YKQO3WHHPWAV","download_json":"https://pith.science/pith/RUYIDA7FUGGDF7YKQO3WHHPWAV.json","view_paper":"https://pith.science/paper/RUYIDA7F","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2210.14348&json=true","fetch_graph":"https://pith.science/api/pith-number/RUYIDA7FUGGDF7YKQO3WHHPWAV/graph.json","fetch_events":"https://pith.science/api/pith-number/RUYIDA7FUGGDF7YKQO3WHHPWAV/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/RUYIDA7FUGGDF7YKQO3WHHPWAV/action/timestamp_anchor","attest_storage":"https://pith.science/pith/RUYIDA7FUGGDF7YKQO3WHHPWAV/action/storage_attestation","attest_author":"https://pith.science/pith/RUYIDA7FUGGDF7YKQO3WHHPWAV/action/author_attestation","sign_citation":"https://pith.science/pith/RUYIDA7FUGGDF7YKQO3WHHPWAV/action/citation_signature","submit_replication":"https://pith.science/pith/RUYIDA7FUGGDF7YKQO3WHHPWAV/action/replication_record"}},"created_at":"2026-07-05T06:32:02.076650+00:00","updated_at":"2026-07-05T06:32:02.076650+00:00"}