{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:CQEWXJEFV7EE3VHGYWQJ624URW","short_pith_number":"pith:CQEWXJEF","canonical_record":{"source":{"id":"2605.20521","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-05-19T21:43:22Z","cross_cats_sorted":["cs.CR"],"title_canon_sha256":"6579a65668032e9eba9ee9664e5583cb3f220e6065b2ff8aeea35e41987dcad2","abstract_canon_sha256":"bf5d20df73ab1845605320d9141142eb12c52bf64a1d30f1b8bec6e4732cc512"},"schema_version":"1.0"},"canonical_sha256":"14096ba485afc84dd4e6c5a09f6b948dbd9e2bcd255ac256cc352b4cecccb0c5","source":{"kind":"arxiv","id":"2605.20521","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.20521","created_at":"2026-05-21T01:04:40Z"},{"alias_kind":"arxiv_version","alias_value":"2605.20521v1","created_at":"2026-05-21T01:04:40Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.20521","created_at":"2026-05-21T01:04:40Z"},{"alias_kind":"pith_short_12","alias_value":"CQEWXJEFV7EE","created_at":"2026-05-21T01:04:40Z"},{"alias_kind":"pith_short_16","alias_value":"CQEWXJEFV7EE3VHG","created_at":"2026-05-21T01:04:40Z"},{"alias_kind":"pith_short_8","alias_value":"CQEWXJEF","created_at":"2026-05-21T01:04:40Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:CQEWXJEFV7EE3VHGYWQJ624URW","target":"record","payload":{"canonical_record":{"source":{"id":"2605.20521","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-05-19T21:43:22Z","cross_cats_sorted":["cs.CR"],"title_canon_sha256":"6579a65668032e9eba9ee9664e5583cb3f220e6065b2ff8aeea35e41987dcad2","abstract_canon_sha256":"bf5d20df73ab1845605320d9141142eb12c52bf64a1d30f1b8bec6e4732cc512"},"schema_version":"1.0"},"canonical_sha256":"14096ba485afc84dd4e6c5a09f6b948dbd9e2bcd255ac256cc352b4cecccb0c5","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-21T01:04:40.856465Z","signature_b64":"hp+FvBoQvYv+NrpWyqQuraTW2QQh16hxBylRsYYD/5b7ZntkyFIhIhwBN5LsjMOp4lh67AIkT/rV3UHHfzd8CA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"14096ba485afc84dd4e6c5a09f6b948dbd9e2bcd255ac256cc352b4cecccb0c5","last_reissued_at":"2026-05-21T01:04:40.856000Z","signature_status":"signed_v1","first_computed_at":"2026-05-21T01:04:40.856000Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2605.20521","source_version":1,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-21T01:04:40Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"8vJJ6lS+MNFPa7olGRd6RGXCSou7qU+0K/Z6a3Cy5ydZcVj3tBx6TSR/KEZkkwCCAHB82Dz7/yBF47d/yNaJDQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-26T21:00:41.172912Z"},"content_sha256":"c2b9d529a9c364f4ec7a5f9543058b77fadd9e4c92f933bdaef869ae56e4d5bc","schema_version":"1.0","event_id":"sha256:c2b9d529a9c364f4ec7a5f9543058b77fadd9e4c92f933bdaef869ae56e4d5bc"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:CQEWXJEFV7EE3VHGYWQJ624URW","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"An exponential mechanism based on quadratic approximations for fine-tuning machine learning models with privacy guarantees","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CR"],"primary_cat":"cs.LG","authors_text":"Alberto Bocchinfuso, Christopher Stanley, Hoang Tran, Jiayi Wang, Jorge Ramirez, M. Paul Laiu","submitted_at":"2026-05-19T21:43:22Z","abstract_excerpt":"Fine-tuning adapts a pretrained machine learning model to a small, sensitive dataset, but this process risks memorizing individual new data points, making the model vulnerable to adversaries who seek to extract sensitive information. In this work, we develop a randomized algorithm based on the exponential mechanism for fine-tuning while ensuring differential privacy. Our key idea is to construct a simple utility function that combines a local quadratic approximation of the pretrained model with information from the new dataset. The resulting exponential mechanism admits exact sampling from a m"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.20521","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/2605.20521/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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-21T01:04:40Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"jrRTEY0P+X7MIn11oe+bHfVlBYEWfdqqTysuf4s6VQ7C74aXpN/w0PCTRCZRbD2nCSttXVpoCNORmTbGldcAAg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-26T21:00:41.173325Z"},"content_sha256":"97a6f611ec42b496bb91270c48d69f033cd900103318f018ce904974c3a08ad6","schema_version":"1.0","event_id":"sha256:97a6f611ec42b496bb91270c48d69f033cd900103318f018ce904974c3a08ad6"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/CQEWXJEFV7EE3VHGYWQJ624URW/bundle.json","state_url":"https://pith.science/pith/CQEWXJEFV7EE3VHGYWQJ624URW/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/CQEWXJEFV7EE3VHGYWQJ624URW/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-05-26T21:00:41Z","links":{"resolver":"https://pith.science/pith/CQEWXJEFV7EE3VHGYWQJ624URW","bundle":"https://pith.science/pith/CQEWXJEFV7EE3VHGYWQJ624URW/bundle.json","state":"https://pith.science/pith/CQEWXJEFV7EE3VHGYWQJ624URW/state.json","well_known_bundle":"https://pith.science/.well-known/pith/CQEWXJEFV7EE3VHGYWQJ624URW/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:CQEWXJEFV7EE3VHGYWQJ624URW","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"bf5d20df73ab1845605320d9141142eb12c52bf64a1d30f1b8bec6e4732cc512","cross_cats_sorted":["cs.CR"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-05-19T21:43:22Z","title_canon_sha256":"6579a65668032e9eba9ee9664e5583cb3f220e6065b2ff8aeea35e41987dcad2"},"schema_version":"1.0","source":{"id":"2605.20521","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.20521","created_at":"2026-05-21T01:04:40Z"},{"alias_kind":"arxiv_version","alias_value":"2605.20521v1","created_at":"2026-05-21T01:04:40Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.20521","created_at":"2026-05-21T01:04:40Z"},{"alias_kind":"pith_short_12","alias_value":"CQEWXJEFV7EE","created_at":"2026-05-21T01:04:40Z"},{"alias_kind":"pith_short_16","alias_value":"CQEWXJEFV7EE3VHG","created_at":"2026-05-21T01:04:40Z"},{"alias_kind":"pith_short_8","alias_value":"CQEWXJEF","created_at":"2026-05-21T01:04:40Z"}],"graph_snapshots":[{"event_id":"sha256:97a6f611ec42b496bb91270c48d69f033cd900103318f018ce904974c3a08ad6","target":"graph","created_at":"2026-05-21T01:04:40Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2605.20521/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Fine-tuning adapts a pretrained machine learning model to a small, sensitive dataset, but this process risks memorizing individual new data points, making the model vulnerable to adversaries who seek to extract sensitive information. In this work, we develop a randomized algorithm based on the exponential mechanism for fine-tuning while ensuring differential privacy. Our key idea is to construct a simple utility function that combines a local quadratic approximation of the pretrained model with information from the new dataset. The resulting exponential mechanism admits exact sampling from a m","authors_text":"Alberto Bocchinfuso, Christopher Stanley, Hoang Tran, Jiayi Wang, Jorge Ramirez, M. Paul Laiu","cross_cats":["cs.CR"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-05-19T21:43:22Z","title":"An exponential mechanism based on quadratic approximations for fine-tuning machine learning models with privacy guarantees"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.20521","kind":"arxiv","version":1},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:c2b9d529a9c364f4ec7a5f9543058b77fadd9e4c92f933bdaef869ae56e4d5bc","target":"record","created_at":"2026-05-21T01:04:40Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"bf5d20df73ab1845605320d9141142eb12c52bf64a1d30f1b8bec6e4732cc512","cross_cats_sorted":["cs.CR"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-05-19T21:43:22Z","title_canon_sha256":"6579a65668032e9eba9ee9664e5583cb3f220e6065b2ff8aeea35e41987dcad2"},"schema_version":"1.0","source":{"id":"2605.20521","kind":"arxiv","version":1}},"canonical_sha256":"14096ba485afc84dd4e6c5a09f6b948dbd9e2bcd255ac256cc352b4cecccb0c5","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"14096ba485afc84dd4e6c5a09f6b948dbd9e2bcd255ac256cc352b4cecccb0c5","first_computed_at":"2026-05-21T01:04:40.856000Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-21T01:04:40.856000Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"hp+FvBoQvYv+NrpWyqQuraTW2QQh16hxBylRsYYD/5b7ZntkyFIhIhwBN5LsjMOp4lh67AIkT/rV3UHHfzd8CA==","signature_status":"signed_v1","signed_at":"2026-05-21T01:04:40.856465Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.20521","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:c2b9d529a9c364f4ec7a5f9543058b77fadd9e4c92f933bdaef869ae56e4d5bc","sha256:97a6f611ec42b496bb91270c48d69f033cd900103318f018ce904974c3a08ad6"],"state_sha256":"337647dcf83328a62d981a5b6b4f881223be37c932c01ce00964a6f20405910b"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"uaIancQqMZ2YekICI8ehtOEqqHDZyhsn7gw5DQCnnGUOuqmEdN4p9wUs/mSp9oedMbxULDOSLA9w7vY/O31kAg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-05-26T21:00:41.177723Z","bundle_sha256":"f1ce85dc0945cc8ecf3eda60f90dce8baff5a0c0e868212689fbf3da5be93240"}}