{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:NWJ37X2HDR54FSTVKYPXWXDXK6","short_pith_number":"pith:NWJ37X2H","schema_version":"1.0","canonical_sha256":"6d93bfdf471c7bc2ca75561f7b5c77578c0a715a6595b4205aaa076ed32ac1a8","source":{"kind":"arxiv","id":"2405.14569","version":3},"attestation_state":"computed","paper":{"title":"PrivCirNet: Efficient Private Inference via Block Circulant Transformation","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CR","authors_text":"Lemeng Wu, Meng Li, Runsheng Wang, Tianshi Xu","submitted_at":"2024-05-23T13:44:48Z","abstract_excerpt":"Homomorphic encryption (HE)-based deep neural network (DNN) inference protects data and model privacy but suffers from significant computation overhead. We observe transforming the DNN weights into circulant matrices converts general matrix-vector multiplications into HE-friendly 1-dimensional convolutions, drastically reducing the HE computation cost. Hence, in this paper, we propose \\method, a protocol/network co-optimization framework based on block circulant transformation. At the protocol level, PrivCirNet customizes the HE encoding algorithm that is fully compatible with the block circul"},"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":"2405.14569","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.CR","submitted_at":"2024-05-23T13:44:48Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"64da9f6b2e0b5b63ecb1337606ad3976272ef4acc6794edeae6523e493c88b73","abstract_canon_sha256":"8c0c851fdccc26b898ab42358dfa145401a6ae186da232f1fde971f161eab813"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:27:22.131527Z","signature_b64":"LWCurxBBp+TIpU5RL4N676x5pHPNgJKLSJQal8nGFZHelrV9w0st8NRuCu65yxlayF+qa/uptuYnzWW5sSrdAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"6d93bfdf471c7bc2ca75561f7b5c77578c0a715a6595b4205aaa076ed32ac1a8","last_reissued_at":"2026-07-05T09:27:22.131093Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:27:22.131093Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"PrivCirNet: Efficient Private Inference via Block Circulant Transformation","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CR","authors_text":"Lemeng Wu, Meng Li, Runsheng Wang, Tianshi Xu","submitted_at":"2024-05-23T13:44:48Z","abstract_excerpt":"Homomorphic encryption (HE)-based deep neural network (DNN) inference protects data and model privacy but suffers from significant computation overhead. We observe transforming the DNN weights into circulant matrices converts general matrix-vector multiplications into HE-friendly 1-dimensional convolutions, drastically reducing the HE computation cost. Hence, in this paper, we propose \\method, a protocol/network co-optimization framework based on block circulant transformation. At the protocol level, PrivCirNet customizes the HE encoding algorithm that is fully compatible with the block circul"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2405.14569","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/2405.14569/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":"2405.14569","created_at":"2026-07-05T09:27:22.131147+00:00"},{"alias_kind":"arxiv_version","alias_value":"2405.14569v3","created_at":"2026-07-05T09:27:22.131147+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2405.14569","created_at":"2026-07-05T09:27:22.131147+00:00"},{"alias_kind":"pith_short_12","alias_value":"NWJ37X2HDR54","created_at":"2026-07-05T09:27:22.131147+00:00"},{"alias_kind":"pith_short_16","alias_value":"NWJ37X2HDR54FSTV","created_at":"2026-07-05T09:27:22.131147+00:00"},{"alias_kind":"pith_short_8","alias_value":"NWJ37X2H","created_at":"2026-07-05T09:27:22.131147+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2507.14519","citing_title":"Towards Efficient Privacy-Preserving Machine Learning: A Systematic Review from Protocol, Model, and System Perspectives","ref_index":186,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/NWJ37X2HDR54FSTVKYPXWXDXK6","json":"https://pith.science/pith/NWJ37X2HDR54FSTVKYPXWXDXK6.json","graph_json":"https://pith.science/api/pith-number/NWJ37X2HDR54FSTVKYPXWXDXK6/graph.json","events_json":"https://pith.science/api/pith-number/NWJ37X2HDR54FSTVKYPXWXDXK6/events.json","paper":"https://pith.science/paper/NWJ37X2H"},"agent_actions":{"view_html":"https://pith.science/pith/NWJ37X2HDR54FSTVKYPXWXDXK6","download_json":"https://pith.science/pith/NWJ37X2HDR54FSTVKYPXWXDXK6.json","view_paper":"https://pith.science/paper/NWJ37X2H","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2405.14569&json=true","fetch_graph":"https://pith.science/api/pith-number/NWJ37X2HDR54FSTVKYPXWXDXK6/graph.json","fetch_events":"https://pith.science/api/pith-number/NWJ37X2HDR54FSTVKYPXWXDXK6/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/NWJ37X2HDR54FSTVKYPXWXDXK6/action/timestamp_anchor","attest_storage":"https://pith.science/pith/NWJ37X2HDR54FSTVKYPXWXDXK6/action/storage_attestation","attest_author":"https://pith.science/pith/NWJ37X2HDR54FSTVKYPXWXDXK6/action/author_attestation","sign_citation":"https://pith.science/pith/NWJ37X2HDR54FSTVKYPXWXDXK6/action/citation_signature","submit_replication":"https://pith.science/pith/NWJ37X2HDR54FSTVKYPXWXDXK6/action/replication_record"}},"created_at":"2026-07-05T09:27:22.131147+00:00","updated_at":"2026-07-05T09:27:22.131147+00:00"}