{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:UOMCA7G6QLL3W2S4DTE4EMLRQX","short_pith_number":"pith:UOMCA7G6","schema_version":"1.0","canonical_sha256":"a398207cde82d7bb6a5c1cc9c2317185e904c9b13cacaf87c59d3c1724b8f72e","source":{"kind":"arxiv","id":"2106.04590","version":2},"attestation_state":"computed","paper":{"title":"PEARL: Data Synthesis via Private Embeddings and Adversarial Reconstruction Learning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CR"],"primary_cat":"cs.LG","authors_text":"Michihiko Ueno, Seng Pei Liew, Tsubasa Takahashi","submitted_at":"2021-06-08T18:00:01Z","abstract_excerpt":"We propose a new framework of synthesizing data using deep generative models in a differentially private manner. Within our framework, sensitive data are sanitized with rigorous privacy guarantees in a one-shot fashion, such that training deep generative models is possible without re-using the original data. Hence, no extra privacy costs or model constraints are incurred, in contrast to popular approaches such as Differentially Private Stochastic Gradient Descent (DP-SGD), which, among other issues, causes degradation in privacy guarantees as the training iteration increases. We demonstrate 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":"2106.04590","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2021-06-08T18:00:01Z","cross_cats_sorted":["cs.CR"],"title_canon_sha256":"23892c79065723442a5790a7081d5029a29e029e9eb0091671ddcc6f60adda93","abstract_canon_sha256":"e8936e8373c20d99d70b0d78d2eeade6e3853f1c7a15c7dc81434b71bfb95552"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T04:02:54.205872Z","signature_b64":"/Y0mMlrt+dmrvIHBjNHfr6rVWOEvx/vK7iTjpT7wk/mQmelPR8lwVO1tKOtGTNxeHy9/czrMHk6g/zNwN6VKDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a398207cde82d7bb6a5c1cc9c2317185e904c9b13cacaf87c59d3c1724b8f72e","last_reissued_at":"2026-07-05T04:02:54.205365Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T04:02:54.205365Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"PEARL: Data Synthesis via Private Embeddings and Adversarial Reconstruction Learning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CR"],"primary_cat":"cs.LG","authors_text":"Michihiko Ueno, Seng Pei Liew, Tsubasa Takahashi","submitted_at":"2021-06-08T18:00:01Z","abstract_excerpt":"We propose a new framework of synthesizing data using deep generative models in a differentially private manner. Within our framework, sensitive data are sanitized with rigorous privacy guarantees in a one-shot fashion, such that training deep generative models is possible without re-using the original data. Hence, no extra privacy costs or model constraints are incurred, in contrast to popular approaches such as Differentially Private Stochastic Gradient Descent (DP-SGD), which, among other issues, causes degradation in privacy guarantees as the training iteration increases. We demonstrate a "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2106.04590","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/2106.04590/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":"2106.04590","created_at":"2026-07-05T04:02:54.205433+00:00"},{"alias_kind":"arxiv_version","alias_value":"2106.04590v2","created_at":"2026-07-05T04:02:54.205433+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2106.04590","created_at":"2026-07-05T04:02:54.205433+00:00"},{"alias_kind":"pith_short_12","alias_value":"UOMCA7G6QLL3","created_at":"2026-07-05T04:02:54.205433+00:00"},{"alias_kind":"pith_short_16","alias_value":"UOMCA7G6QLL3W2S4","created_at":"2026-07-05T04:02:54.205433+00:00"},{"alias_kind":"pith_short_8","alias_value":"UOMCA7G6","created_at":"2026-07-05T04:02:54.205433+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2412.09842","citing_title":"Leveraging Programmatically Generated Synthetic Data for Differentially Private Diffusion Training","ref_index":20,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/UOMCA7G6QLL3W2S4DTE4EMLRQX","json":"https://pith.science/pith/UOMCA7G6QLL3W2S4DTE4EMLRQX.json","graph_json":"https://pith.science/api/pith-number/UOMCA7G6QLL3W2S4DTE4EMLRQX/graph.json","events_json":"https://pith.science/api/pith-number/UOMCA7G6QLL3W2S4DTE4EMLRQX/events.json","paper":"https://pith.science/paper/UOMCA7G6"},"agent_actions":{"view_html":"https://pith.science/pith/UOMCA7G6QLL3W2S4DTE4EMLRQX","download_json":"https://pith.science/pith/UOMCA7G6QLL3W2S4DTE4EMLRQX.json","view_paper":"https://pith.science/paper/UOMCA7G6","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2106.04590&json=true","fetch_graph":"https://pith.science/api/pith-number/UOMCA7G6QLL3W2S4DTE4EMLRQX/graph.json","fetch_events":"https://pith.science/api/pith-number/UOMCA7G6QLL3W2S4DTE4EMLRQX/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/UOMCA7G6QLL3W2S4DTE4EMLRQX/action/timestamp_anchor","attest_storage":"https://pith.science/pith/UOMCA7G6QLL3W2S4DTE4EMLRQX/action/storage_attestation","attest_author":"https://pith.science/pith/UOMCA7G6QLL3W2S4DTE4EMLRQX/action/author_attestation","sign_citation":"https://pith.science/pith/UOMCA7G6QLL3W2S4DTE4EMLRQX/action/citation_signature","submit_replication":"https://pith.science/pith/UOMCA7G6QLL3W2S4DTE4EMLRQX/action/replication_record"}},"created_at":"2026-07-05T04:02:54.205433+00:00","updated_at":"2026-07-05T04:02:54.205433+00:00"}