{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:3IPOIBR63PYGD3GRVK4P3BNNGF","short_pith_number":"pith:3IPOIBR6","schema_version":"1.0","canonical_sha256":"da1ee4063edbf061ecd1aab8fd85ad3152df62429786a6c3b14bcda97ac4f73d","source":{"kind":"arxiv","id":"2605.23641","version":1},"attestation_state":"computed","paper":{"title":"Kernel-Based ReLU Approximation for Homomorphic Encryption-Compatible Privacy-preserving Deep Learning Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CR","authors_text":"Dimitra Papatsaroucha, Dimitrios Sygletos, Evangelos K. Markakis, Ilias Politis, Marios Choudetsanakis","submitted_at":"2026-05-22T13:54:37Z","abstract_excerpt":"As privacy concerns in AI technologies continue to grow, Homomorphic Encryption (HE) offers a way to perform computations on encrypted data without the need of decryption during operations. However, HE is limited to addition and multiplication, making non-linear functions incompatible in their original form. This limitation has become more critical with the widespread use of Large Language Models (LLMs), where the non-linearity of activation functions such as the Rectified Linear Unit (ReLU) poses challenges for deployment in privacy-preserving Natural Language Processing (NLP) settings. This "},"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":"2605.23641","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2026-05-22T13:54:37Z","cross_cats_sorted":[],"title_canon_sha256":"34f9fb0a08c0e9f055a5a9f7dcb82a3b607966985bb71ba7eae6d32406852f17","abstract_canon_sha256":"0cd31e362e8797c605dded6c70d84f919737c54e69d43d6f9ea354abb75d9939"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-25T02:02:23.474698Z","signature_b64":"P54uOEtO3vAkEjunWjjZdubIk74P6uwUwx066rYZGw2HVswsNGtO3fb11oYkjq5CX1KNsF1Hh7SDPo4xyWX0AA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"da1ee4063edbf061ecd1aab8fd85ad3152df62429786a6c3b14bcda97ac4f73d","last_reissued_at":"2026-05-25T02:02:23.473866Z","signature_status":"signed_v1","first_computed_at":"2026-05-25T02:02:23.473866Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Kernel-Based ReLU Approximation for Homomorphic Encryption-Compatible Privacy-preserving Deep Learning Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CR","authors_text":"Dimitra Papatsaroucha, Dimitrios Sygletos, Evangelos K. Markakis, Ilias Politis, Marios Choudetsanakis","submitted_at":"2026-05-22T13:54:37Z","abstract_excerpt":"As privacy concerns in AI technologies continue to grow, Homomorphic Encryption (HE) offers a way to perform computations on encrypted data without the need of decryption during operations. However, HE is limited to addition and multiplication, making non-linear functions incompatible in their original form. This limitation has become more critical with the widespread use of Large Language Models (LLMs), where the non-linearity of activation functions such as the Rectified Linear Unit (ReLU) poses challenges for deployment in privacy-preserving Natural Language Processing (NLP) settings. This "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.23641","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/2605.23641/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":"2605.23641","created_at":"2026-05-25T02:02:23.473976+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.23641v1","created_at":"2026-05-25T02:02:23.473976+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.23641","created_at":"2026-05-25T02:02:23.473976+00:00"},{"alias_kind":"pith_short_12","alias_value":"3IPOIBR63PYG","created_at":"2026-05-25T02:02:23.473976+00:00"},{"alias_kind":"pith_short_16","alias_value":"3IPOIBR63PYGD3GR","created_at":"2026-05-25T02:02:23.473976+00:00"},{"alias_kind":"pith_short_8","alias_value":"3IPOIBR6","created_at":"2026-05-25T02:02:23.473976+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/3IPOIBR63PYGD3GRVK4P3BNNGF","json":"https://pith.science/pith/3IPOIBR63PYGD3GRVK4P3BNNGF.json","graph_json":"https://pith.science/api/pith-number/3IPOIBR63PYGD3GRVK4P3BNNGF/graph.json","events_json":"https://pith.science/api/pith-number/3IPOIBR63PYGD3GRVK4P3BNNGF/events.json","paper":"https://pith.science/paper/3IPOIBR6"},"agent_actions":{"view_html":"https://pith.science/pith/3IPOIBR63PYGD3GRVK4P3BNNGF","download_json":"https://pith.science/pith/3IPOIBR63PYGD3GRVK4P3BNNGF.json","view_paper":"https://pith.science/paper/3IPOIBR6","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.23641&json=true","fetch_graph":"https://pith.science/api/pith-number/3IPOIBR63PYGD3GRVK4P3BNNGF/graph.json","fetch_events":"https://pith.science/api/pith-number/3IPOIBR63PYGD3GRVK4P3BNNGF/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/3IPOIBR63PYGD3GRVK4P3BNNGF/action/timestamp_anchor","attest_storage":"https://pith.science/pith/3IPOIBR63PYGD3GRVK4P3BNNGF/action/storage_attestation","attest_author":"https://pith.science/pith/3IPOIBR63PYGD3GRVK4P3BNNGF/action/author_attestation","sign_citation":"https://pith.science/pith/3IPOIBR63PYGD3GRVK4P3BNNGF/action/citation_signature","submit_replication":"https://pith.science/pith/3IPOIBR63PYGD3GRVK4P3BNNGF/action/replication_record"}},"created_at":"2026-05-25T02:02:23.473976+00:00","updated_at":"2026-05-25T02:02:23.473976+00:00"}