{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:WLNDXX4MBDI5ENGE4JQN6KDHMM","short_pith_number":"pith:WLNDXX4M","schema_version":"1.0","canonical_sha256":"b2da3bdf8c08d1d234c4e260df28676330f3779f01fc610c3ada73323c16824f","source":{"kind":"arxiv","id":"2507.16872","version":1},"attestation_state":"computed","paper":{"title":"CompLeak: Deep Learning Model Compression Exacerbates Privacy Leakage","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CR","authors_text":"Anmin Fu, Boyu Kuang, Hongsheng Hu, Na Li, Yansong Gao","submitted_at":"2025-07-22T08:02:46Z","abstract_excerpt":"Model compression is crucial for minimizing memory storage and accelerating inference in deep learning (DL) models, including recent foundation models like large language models (LLMs). Users can access different compressed model versions according to their resources and budget. However, while existing compression operations primarily focus on optimizing the trade-off between resource efficiency and model performance, the privacy risks introduced by compression remain overlooked and insufficiently understood.\n  In this work, through the lens of membership inference attack (MIA), we propose Com"},"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":"2507.16872","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CR","submitted_at":"2025-07-22T08:02:46Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"b569023e4fb21bbb404b44c06c030a15c7010fc4932f7b90873f3ab3c1a71d65","abstract_canon_sha256":"1202fef459b5d975bf53871bca8db64798bfa66ca3225aea79ba51af40636283"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:41:57.514197Z","signature_b64":"4MjTO1J5xykBLXj+7vRlZwlCOHo6TUXZzyS6k70R/pxF/pUmtFcmJGXK1iCgfQOGocwYgq1QairHiYkRoe/dCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b2da3bdf8c08d1d234c4e260df28676330f3779f01fc610c3ada73323c16824f","last_reissued_at":"2026-07-05T11:41:57.513632Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:41:57.513632Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"CompLeak: Deep Learning Model Compression Exacerbates Privacy Leakage","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CR","authors_text":"Anmin Fu, Boyu Kuang, Hongsheng Hu, Na Li, Yansong Gao","submitted_at":"2025-07-22T08:02:46Z","abstract_excerpt":"Model compression is crucial for minimizing memory storage and accelerating inference in deep learning (DL) models, including recent foundation models like large language models (LLMs). Users can access different compressed model versions according to their resources and budget. However, while existing compression operations primarily focus on optimizing the trade-off between resource efficiency and model performance, the privacy risks introduced by compression remain overlooked and insufficiently understood.\n  In this work, through the lens of membership inference attack (MIA), we propose Com"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2507.16872","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/2507.16872/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":"2507.16872","created_at":"2026-07-05T11:41:57.513710+00:00"},{"alias_kind":"arxiv_version","alias_value":"2507.16872v1","created_at":"2026-07-05T11:41:57.513710+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2507.16872","created_at":"2026-07-05T11:41:57.513710+00:00"},{"alias_kind":"pith_short_12","alias_value":"WLNDXX4MBDI5","created_at":"2026-07-05T11:41:57.513710+00:00"},{"alias_kind":"pith_short_16","alias_value":"WLNDXX4MBDI5ENGE","created_at":"2026-07-05T11:41:57.513710+00:00"},{"alias_kind":"pith_short_8","alias_value":"WLNDXX4M","created_at":"2026-07-05T11:41:57.513710+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/WLNDXX4MBDI5ENGE4JQN6KDHMM","json":"https://pith.science/pith/WLNDXX4MBDI5ENGE4JQN6KDHMM.json","graph_json":"https://pith.science/api/pith-number/WLNDXX4MBDI5ENGE4JQN6KDHMM/graph.json","events_json":"https://pith.science/api/pith-number/WLNDXX4MBDI5ENGE4JQN6KDHMM/events.json","paper":"https://pith.science/paper/WLNDXX4M"},"agent_actions":{"view_html":"https://pith.science/pith/WLNDXX4MBDI5ENGE4JQN6KDHMM","download_json":"https://pith.science/pith/WLNDXX4MBDI5ENGE4JQN6KDHMM.json","view_paper":"https://pith.science/paper/WLNDXX4M","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2507.16872&json=true","fetch_graph":"https://pith.science/api/pith-number/WLNDXX4MBDI5ENGE4JQN6KDHMM/graph.json","fetch_events":"https://pith.science/api/pith-number/WLNDXX4MBDI5ENGE4JQN6KDHMM/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/WLNDXX4MBDI5ENGE4JQN6KDHMM/action/timestamp_anchor","attest_storage":"https://pith.science/pith/WLNDXX4MBDI5ENGE4JQN6KDHMM/action/storage_attestation","attest_author":"https://pith.science/pith/WLNDXX4MBDI5ENGE4JQN6KDHMM/action/author_attestation","sign_citation":"https://pith.science/pith/WLNDXX4MBDI5ENGE4JQN6KDHMM/action/citation_signature","submit_replication":"https://pith.science/pith/WLNDXX4MBDI5ENGE4JQN6KDHMM/action/replication_record"}},"created_at":"2026-07-05T11:41:57.513710+00:00","updated_at":"2026-07-05T11:41:57.513710+00:00"}