{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2022:HWVO3YJKG6CK2E6523U5V2SP2W","short_pith_number":"pith:HWVO3YJK","canonical_record":{"source":{"id":"2211.04686","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2022-11-09T05:18:08Z","cross_cats_sorted":["cs.CR"],"title_canon_sha256":"9d21e629896f9abc7b582de2e79c8872d02b9a92330f4d9e831e6457ef3705c3","abstract_canon_sha256":"5022696f8824968757dfe4317d4ee4d589afda99cb21aa31d7161c76854d0e6c"},"schema_version":"1.0"},"canonical_sha256":"3daaede12a3784ad13ddd6e9daea4fd5ad21730068f20acc405aeb3199d89c9d","source":{"kind":"arxiv","id":"2211.04686","version":3},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2211.04686","created_at":"2026-07-05T07:16:34Z"},{"alias_kind":"arxiv_version","alias_value":"2211.04686v3","created_at":"2026-07-05T07:16:34Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2211.04686","created_at":"2026-07-05T07:16:34Z"},{"alias_kind":"pith_short_12","alias_value":"HWVO3YJKG6CK","created_at":"2026-07-05T07:16:34Z"},{"alias_kind":"pith_short_16","alias_value":"HWVO3YJKG6CK2E65","created_at":"2026-07-05T07:16:34Z"},{"alias_kind":"pith_short_8","alias_value":"HWVO3YJK","created_at":"2026-07-05T07:16:34Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2022:HWVO3YJKG6CK2E6523U5V2SP2W","target":"record","payload":{"canonical_record":{"source":{"id":"2211.04686","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2022-11-09T05:18:08Z","cross_cats_sorted":["cs.CR"],"title_canon_sha256":"9d21e629896f9abc7b582de2e79c8872d02b9a92330f4d9e831e6457ef3705c3","abstract_canon_sha256":"5022696f8824968757dfe4317d4ee4d589afda99cb21aa31d7161c76854d0e6c"},"schema_version":"1.0"},"canonical_sha256":"3daaede12a3784ad13ddd6e9daea4fd5ad21730068f20acc405aeb3199d89c9d","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:16:34.769032Z","signature_b64":"SOISol69BLc2v/WjnoC2md7QnwH3GkwfNESUgoTr56mzOD9i8CGtE7GK50XvOQxkjqKVynkswTROxhcfhWAnDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"3daaede12a3784ad13ddd6e9daea4fd5ad21730068f20acc405aeb3199d89c9d","last_reissued_at":"2026-07-05T07:16:34.768532Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:16:34.768532Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2211.04686","source_version":3,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T07:16:34Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"prrjPCZfkRIoxN6kwwaYsgEfLmwEVTE5lMbUBUVDORBep0xFx8u8W/+poOiK3FbJZ+iTD2NErs9MqsAkr5IfCQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-09T23:46:36.836591Z"},"content_sha256":"51fa11cef26e2d3ac03f0b15b7a033df2b677dbf6415a847525adb4d6efda3b7","schema_version":"1.0","event_id":"sha256:51fa11cef26e2d3ac03f0b15b7a033df2b677dbf6415a847525adb4d6efda3b7"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2022:HWVO3YJKG6CK2E6523U5V2SP2W","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Directional Privacy for Deep Learning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CR"],"primary_cat":"cs.LG","authors_text":"Annabelle McIver, Mark Dras, Natasha Fernandes, Pedro Faustini, Shakila Tonni","submitted_at":"2022-11-09T05:18:08Z","abstract_excerpt":"Differentially Private Stochastic Gradient Descent (DP-SGD) is a key method for applying privacy in the training of deep learning models. It applies isotropic Gaussian noise to gradients during training, which can perturb these gradients in any direction, damaging utility. Metric DP, however, can provide alternative mechanisms based on arbitrary metrics that might be more suitable for preserving utility. In this paper, we apply \\textit{directional privacy}, via a mechanism based on the von Mises-Fisher (VMF) distribution, to perturb gradients in terms of \\textit{angular distance} so that gradi"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2211.04686","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/2211.04686/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":"2026-07-05T07:16:34Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"v+uKh6DVUMlbwkglhy32k6d2X7Q/O4tYxT2yVjmlTozKKEmdcGPI9suGHl2a9O/tSYFvhhqMZId2XPb+SUgfBw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-09T23:46:36.837071Z"},"content_sha256":"ae830ff37c579f3558a27aec5dec6c13c99082ba298aeec4b2900a8c99923ff3","schema_version":"1.0","event_id":"sha256:ae830ff37c579f3558a27aec5dec6c13c99082ba298aeec4b2900a8c99923ff3"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/HWVO3YJKG6CK2E6523U5V2SP2W/bundle.json","state_url":"https://pith.science/pith/HWVO3YJKG6CK2E6523U5V2SP2W/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/HWVO3YJKG6CK2E6523U5V2SP2W/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":"2026-08-09T23:46:36Z","links":{"resolver":"https://pith.science/pith/HWVO3YJKG6CK2E6523U5V2SP2W","bundle":"https://pith.science/pith/HWVO3YJKG6CK2E6523U5V2SP2W/bundle.json","state":"https://pith.science/pith/HWVO3YJKG6CK2E6523U5V2SP2W/state.json","well_known_bundle":"https://pith.science/.well-known/pith/HWVO3YJKG6CK2E6523U5V2SP2W/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2022:HWVO3YJKG6CK2E6523U5V2SP2W","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":"5022696f8824968757dfe4317d4ee4d589afda99cb21aa31d7161c76854d0e6c","cross_cats_sorted":["cs.CR"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2022-11-09T05:18:08Z","title_canon_sha256":"9d21e629896f9abc7b582de2e79c8872d02b9a92330f4d9e831e6457ef3705c3"},"schema_version":"1.0","source":{"id":"2211.04686","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2211.04686","created_at":"2026-07-05T07:16:34Z"},{"alias_kind":"arxiv_version","alias_value":"2211.04686v3","created_at":"2026-07-05T07:16:34Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2211.04686","created_at":"2026-07-05T07:16:34Z"},{"alias_kind":"pith_short_12","alias_value":"HWVO3YJKG6CK","created_at":"2026-07-05T07:16:34Z"},{"alias_kind":"pith_short_16","alias_value":"HWVO3YJKG6CK2E65","created_at":"2026-07-05T07:16:34Z"},{"alias_kind":"pith_short_8","alias_value":"HWVO3YJK","created_at":"2026-07-05T07:16:34Z"}],"graph_snapshots":[{"event_id":"sha256:ae830ff37c579f3558a27aec5dec6c13c99082ba298aeec4b2900a8c99923ff3","target":"graph","created_at":"2026-07-05T07:16:34Z","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/2211.04686/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Differentially Private Stochastic Gradient Descent (DP-SGD) is a key method for applying privacy in the training of deep learning models. It applies isotropic Gaussian noise to gradients during training, which can perturb these gradients in any direction, damaging utility. Metric DP, however, can provide alternative mechanisms based on arbitrary metrics that might be more suitable for preserving utility. In this paper, we apply \\textit{directional privacy}, via a mechanism based on the von Mises-Fisher (VMF) distribution, to perturb gradients in terms of \\textit{angular distance} so that gradi","authors_text":"Annabelle McIver, Mark Dras, Natasha Fernandes, Pedro Faustini, Shakila Tonni","cross_cats":["cs.CR"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2022-11-09T05:18:08Z","title":"Directional Privacy for Deep Learning"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2211.04686","kind":"arxiv","version":3},"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":"sha256:51fa11cef26e2d3ac03f0b15b7a033df2b677dbf6415a847525adb4d6efda3b7","target":"record","created_at":"2026-07-05T07:16:34Z","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":"5022696f8824968757dfe4317d4ee4d589afda99cb21aa31d7161c76854d0e6c","cross_cats_sorted":["cs.CR"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2022-11-09T05:18:08Z","title_canon_sha256":"9d21e629896f9abc7b582de2e79c8872d02b9a92330f4d9e831e6457ef3705c3"},"schema_version":"1.0","source":{"id":"2211.04686","kind":"arxiv","version":3}},"canonical_sha256":"3daaede12a3784ad13ddd6e9daea4fd5ad21730068f20acc405aeb3199d89c9d","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"3daaede12a3784ad13ddd6e9daea4fd5ad21730068f20acc405aeb3199d89c9d","first_computed_at":"2026-07-05T07:16:34.768532Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T07:16:34.768532Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"SOISol69BLc2v/WjnoC2md7QnwH3GkwfNESUgoTr56mzOD9i8CGtE7GK50XvOQxkjqKVynkswTROxhcfhWAnDQ==","signature_status":"signed_v1","signed_at":"2026-07-05T07:16:34.769032Z","signed_message":"canonical_sha256_bytes"},"source_id":"2211.04686","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:51fa11cef26e2d3ac03f0b15b7a033df2b677dbf6415a847525adb4d6efda3b7","sha256:ae830ff37c579f3558a27aec5dec6c13c99082ba298aeec4b2900a8c99923ff3"],"state_sha256":"8a7c8d41e1c11fda530b16f8f17764d320f71e8cdd495e91d8a9250e09e0b4a4"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"E2wI1ZgaW0x4wRaeDLqhQullEgrJBCmKpOE2SiS8qT2pbJnd0FpeHHgEKlUXT4QyiAwnv1wqkKcqjbPjstRFBQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-09T23:46:36.840794Z","bundle_sha256":"65e8cfa4c657bba156337446893868ce06fc116925830476998d527e7773125b"}}