{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:CJ7NXYKPYSVVJYAOBYG6CQPMOT","short_pith_number":"pith:CJ7NXYKP","schema_version":"1.0","canonical_sha256":"127edbe14fc4ab54e00e0e0de141ec74eb48df5665daae130159cd43f4e20441","source":{"kind":"arxiv","id":"2506.18150","version":4},"attestation_state":"computed","paper":{"title":"HE-LRM: Encrypted Deep Learning Recommendation Models using Fully Homomorphic Encryption","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CR","authors_text":"Austin Ebel, Brandon Reagen, Gabrielle De Micheli, Karthik Garimella","submitted_at":"2025-06-22T19:40:04Z","abstract_excerpt":"Fully Homomorphic Encryption (FHE) enables computation directly on encrypted data and privacy-preserving neural inference in the cloud. Existing solutions focus on models with dense inputs (e.g., CNNs and MLPs). Recommendation models (e.g., DLRM) pose a different challenge: sparse categorical inputs require private lookups into large embedding tables, which must be implemented using FHE's restrictive operators. Naive lookups incur significant communication and memory costs; prior work proposes compressing embedding tables at the expense of introducing large server-side compute costs (i.e., ind"},"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":"2506.18150","kind":"arxiv","version":4},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2025-06-22T19:40:04Z","cross_cats_sorted":[],"title_canon_sha256":"70068634c5fe5e7901a9fc41d66db8ae6a77c0635f311258dc07a3e0ac8cb9fc","abstract_canon_sha256":"f9c791ef132ed55fb7d3739c132d15d9dabacc3dc549a24aa160269dde1582f9"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-21T01:20:33.784251Z","signature_b64":"Lr2FTiE1/VSqcQNAG6SNxuVJvSij+dp0flGdrpfzxRhkmo9ez5ujX6YBliJjQx54me0wMCgqRLWyF44iC9LmCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"127edbe14fc4ab54e00e0e0de141ec74eb48df5665daae130159cd43f4e20441","last_reissued_at":"2026-07-21T01:20:33.783338Z","signature_status":"signed_v1","first_computed_at":"2026-07-21T01:20:33.783338Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"HE-LRM: Encrypted Deep Learning Recommendation Models using Fully Homomorphic Encryption","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CR","authors_text":"Austin Ebel, Brandon Reagen, Gabrielle De Micheli, Karthik Garimella","submitted_at":"2025-06-22T19:40:04Z","abstract_excerpt":"Fully Homomorphic Encryption (FHE) enables computation directly on encrypted data and privacy-preserving neural inference in the cloud. Existing solutions focus on models with dense inputs (e.g., CNNs and MLPs). Recommendation models (e.g., DLRM) pose a different challenge: sparse categorical inputs require private lookups into large embedding tables, which must be implemented using FHE's restrictive operators. Naive lookups incur significant communication and memory costs; prior work proposes compressing embedding tables at the expense of introducing large server-side compute costs (i.e., ind"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.18150","kind":"arxiv","version":4},"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/2506.18150/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":"2506.18150","created_at":"2026-07-21T01:20:33.783762+00:00"},{"alias_kind":"arxiv_version","alias_value":"2506.18150v4","created_at":"2026-07-21T01:20:33.783762+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.18150","created_at":"2026-07-21T01:20:33.783762+00:00"},{"alias_kind":"pith_short_12","alias_value":"CJ7NXYKPYSVV","created_at":"2026-07-21T01:20:33.783762+00:00"},{"alias_kind":"pith_short_16","alias_value":"CJ7NXYKPYSVVJYAO","created_at":"2026-07-21T01:20:33.783762+00:00"},{"alias_kind":"pith_short_8","alias_value":"CJ7NXYKP","created_at":"2026-07-21T01:20:33.783762+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2509.03024","citing_title":"Efficient Privacy-Preserving Recommendation on Sparse Data using Fully Homomorphic Encryption","ref_index":41,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/CJ7NXYKPYSVVJYAOBYG6CQPMOT","json":"https://pith.science/pith/CJ7NXYKPYSVVJYAOBYG6CQPMOT.json","graph_json":"https://pith.science/api/pith-number/CJ7NXYKPYSVVJYAOBYG6CQPMOT/graph.json","events_json":"https://pith.science/api/pith-number/CJ7NXYKPYSVVJYAOBYG6CQPMOT/events.json","paper":"https://pith.science/paper/CJ7NXYKP"},"agent_actions":{"view_html":"https://pith.science/pith/CJ7NXYKPYSVVJYAOBYG6CQPMOT","download_json":"https://pith.science/pith/CJ7NXYKPYSVVJYAOBYG6CQPMOT.json","view_paper":"https://pith.science/paper/CJ7NXYKP","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2506.18150&json=true","fetch_graph":"https://pith.science/api/pith-number/CJ7NXYKPYSVVJYAOBYG6CQPMOT/graph.json","fetch_events":"https://pith.science/api/pith-number/CJ7NXYKPYSVVJYAOBYG6CQPMOT/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/CJ7NXYKPYSVVJYAOBYG6CQPMOT/action/timestamp_anchor","attest_storage":"https://pith.science/pith/CJ7NXYKPYSVVJYAOBYG6CQPMOT/action/storage_attestation","attest_author":"https://pith.science/pith/CJ7NXYKPYSVVJYAOBYG6CQPMOT/action/author_attestation","sign_citation":"https://pith.science/pith/CJ7NXYKPYSVVJYAOBYG6CQPMOT/action/citation_signature","submit_replication":"https://pith.science/pith/CJ7NXYKPYSVVJYAOBYG6CQPMOT/action/replication_record"}},"created_at":"2026-07-21T01:20:33.783762+00:00","updated_at":"2026-07-21T01:20:33.783762+00:00"}