{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:Z5GNBTCY6JELWK7QVLIFWNZCKL","short_pith_number":"pith:Z5GNBTCY","schema_version":"1.0","canonical_sha256":"cf4cd0cc58f248bb2bf0aad05b372252ce6db344ab5ad134dac424e1e770e22b","source":{"kind":"arxiv","id":"2205.02973","version":2},"attestation_state":"computed","paper":{"title":"Large Scale Transfer Learning for Differentially Private Image Classification","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CR","cs.CV"],"primary_cat":"cs.LG","authors_text":"Abhradeep Thakurta, Alexey Kurakin, Ashok Cutkosky, Harsh Mehta","submitted_at":"2022-05-06T01:22:20Z","abstract_excerpt":"Differential Privacy (DP) provides a formal framework for training machine learning models with individual example level privacy. In the field of deep learning, Differentially Private Stochastic Gradient Descent (DP-SGD) has emerged as a popular private training algorithm. Unfortunately, the computational cost of training large-scale models with DP-SGD is substantially higher than non-private training. This is further exacerbated by the fact that increasing the number of parameters leads to larger degradation in utility with DP. In this work, we zoom in on the ImageNet dataset and demonstrate "},"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":"2205.02973","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2022-05-06T01:22:20Z","cross_cats_sorted":["cs.CR","cs.CV"],"title_canon_sha256":"1473a0e9f11eaf6a70b81a0abc37ece8cde6dda0d9073bb2a2aaabc6151f293c","abstract_canon_sha256":"32fd7ada6f7340759b624f7c7626a09028a2007cd0e2d5a85ee6a07f2aab9f66"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T04:25:17.978463Z","signature_b64":"f3a/b65OwZtxQypbid5CSf8I9pPm63sdFFdZpdtvWv0o6D8dp5YUa5VlQlbsOukY106DmZ0JmCdfy07mdeLKCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"cf4cd0cc58f248bb2bf0aad05b372252ce6db344ab5ad134dac424e1e770e22b","last_reissued_at":"2026-07-05T04:25:17.977972Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T04:25:17.977972Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Large Scale Transfer Learning for Differentially Private Image Classification","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CR","cs.CV"],"primary_cat":"cs.LG","authors_text":"Abhradeep Thakurta, Alexey Kurakin, Ashok Cutkosky, Harsh Mehta","submitted_at":"2022-05-06T01:22:20Z","abstract_excerpt":"Differential Privacy (DP) provides a formal framework for training machine learning models with individual example level privacy. In the field of deep learning, Differentially Private Stochastic Gradient Descent (DP-SGD) has emerged as a popular private training algorithm. Unfortunately, the computational cost of training large-scale models with DP-SGD is substantially higher than non-private training. This is further exacerbated by the fact that increasing the number of parameters leads to larger degradation in utility with DP. In this work, we zoom in on the ImageNet dataset and demonstrate "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2205.02973","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/2205.02973/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":"2205.02973","created_at":"2026-07-05T04:25:17.978028+00:00"},{"alias_kind":"arxiv_version","alias_value":"2205.02973v2","created_at":"2026-07-05T04:25:17.978028+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2205.02973","created_at":"2026-07-05T04:25:17.978028+00:00"},{"alias_kind":"pith_short_12","alias_value":"Z5GNBTCY6JEL","created_at":"2026-07-05T04:25:17.978028+00:00"},{"alias_kind":"pith_short_16","alias_value":"Z5GNBTCY6JELWK7Q","created_at":"2026-07-05T04:25:17.978028+00:00"},{"alias_kind":"pith_short_8","alias_value":"Z5GNBTCY","created_at":"2026-07-05T04:25:17.978028+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/Z5GNBTCY6JELWK7QVLIFWNZCKL","json":"https://pith.science/pith/Z5GNBTCY6JELWK7QVLIFWNZCKL.json","graph_json":"https://pith.science/api/pith-number/Z5GNBTCY6JELWK7QVLIFWNZCKL/graph.json","events_json":"https://pith.science/api/pith-number/Z5GNBTCY6JELWK7QVLIFWNZCKL/events.json","paper":"https://pith.science/paper/Z5GNBTCY"},"agent_actions":{"view_html":"https://pith.science/pith/Z5GNBTCY6JELWK7QVLIFWNZCKL","download_json":"https://pith.science/pith/Z5GNBTCY6JELWK7QVLIFWNZCKL.json","view_paper":"https://pith.science/paper/Z5GNBTCY","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2205.02973&json=true","fetch_graph":"https://pith.science/api/pith-number/Z5GNBTCY6JELWK7QVLIFWNZCKL/graph.json","fetch_events":"https://pith.science/api/pith-number/Z5GNBTCY6JELWK7QVLIFWNZCKL/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/Z5GNBTCY6JELWK7QVLIFWNZCKL/action/timestamp_anchor","attest_storage":"https://pith.science/pith/Z5GNBTCY6JELWK7QVLIFWNZCKL/action/storage_attestation","attest_author":"https://pith.science/pith/Z5GNBTCY6JELWK7QVLIFWNZCKL/action/author_attestation","sign_citation":"https://pith.science/pith/Z5GNBTCY6JELWK7QVLIFWNZCKL/action/citation_signature","submit_replication":"https://pith.science/pith/Z5GNBTCY6JELWK7QVLIFWNZCKL/action/replication_record"}},"created_at":"2026-07-05T04:25:17.978028+00:00","updated_at":"2026-07-05T04:25:17.978028+00:00"}