{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:JN72IUBQH2RYBZBBYAW3FBU2BQ","short_pith_number":"pith:JN72IUBQ","schema_version":"1.0","canonical_sha256":"4b7fa450303ea380e421c02db2869a0c1969e7ac5074d13d531f5f33344c472c","source":{"kind":"arxiv","id":"2509.06026","version":1},"attestation_state":"computed","paper":{"title":"DCMI: A Differential Calibration Membership Inference Attack Against Retrieval-Augmented Generation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CR","authors_text":"Shanqing Guo, Xiangtao Meng, Xinyu Gao, Yingkai Dong, Zheng Li","submitted_at":"2025-09-07T11:58:02Z","abstract_excerpt":"While Retrieval-Augmented Generation (RAG) effectively reduces hallucinations by integrating external knowledge bases, it introduces vulnerabilities to membership inference attacks (MIAs), particularly in systems handling sensitive data. Existing MIAs targeting RAG's external databases often rely on model responses but ignore the interference of non-member-retrieved documents on RAG outputs, limiting their effectiveness. To address this, we propose DCMI, a differential calibration MIA that mitigates the negative impact of non-member-retrieved documents. Specifically, DCMI leverages the sensiti"},"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":"2509.06026","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CR","submitted_at":"2025-09-07T11:58:02Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"656acd7122f8f2c9c977d1c01784508c9432b789eeec39a48b57614278609921","abstract_canon_sha256":"1b71e7e168bb197b973f450ed1bae8c8bd82c8238c2e3930aa72d0cc0b1dc6db"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T12:06:28.555845Z","signature_b64":"aBzuv/5JfmhYqQ+rhMnmQwNwZsTFEDJtAJJXz757R1Fz6Ymz1haBxOEmDOIKZ3NNJPmbRx18UD6E4bmtl2XLCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"4b7fa450303ea380e421c02db2869a0c1969e7ac5074d13d531f5f33344c472c","last_reissued_at":"2026-07-05T12:06:28.555344Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T12:06:28.555344Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"DCMI: A Differential Calibration Membership Inference Attack Against Retrieval-Augmented Generation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CR","authors_text":"Shanqing Guo, Xiangtao Meng, Xinyu Gao, Yingkai Dong, Zheng Li","submitted_at":"2025-09-07T11:58:02Z","abstract_excerpt":"While Retrieval-Augmented Generation (RAG) effectively reduces hallucinations by integrating external knowledge bases, it introduces vulnerabilities to membership inference attacks (MIAs), particularly in systems handling sensitive data. Existing MIAs targeting RAG's external databases often rely on model responses but ignore the interference of non-member-retrieved documents on RAG outputs, limiting their effectiveness. To address this, we propose DCMI, a differential calibration MIA that mitigates the negative impact of non-member-retrieved documents. Specifically, DCMI leverages the sensiti"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2509.06026","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/2509.06026/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":"2509.06026","created_at":"2026-07-05T12:06:28.555397+00:00"},{"alias_kind":"arxiv_version","alias_value":"2509.06026v1","created_at":"2026-07-05T12:06:28.555397+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2509.06026","created_at":"2026-07-05T12:06:28.555397+00:00"},{"alias_kind":"pith_short_12","alias_value":"JN72IUBQH2RY","created_at":"2026-07-05T12:06:28.555397+00:00"},{"alias_kind":"pith_short_16","alias_value":"JN72IUBQH2RYBZBB","created_at":"2026-07-05T12:06:28.555397+00:00"},{"alias_kind":"pith_short_8","alias_value":"JN72IUBQ","created_at":"2026-07-05T12:06:28.555397+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/JN72IUBQH2RYBZBBYAW3FBU2BQ","json":"https://pith.science/pith/JN72IUBQH2RYBZBBYAW3FBU2BQ.json","graph_json":"https://pith.science/api/pith-number/JN72IUBQH2RYBZBBYAW3FBU2BQ/graph.json","events_json":"https://pith.science/api/pith-number/JN72IUBQH2RYBZBBYAW3FBU2BQ/events.json","paper":"https://pith.science/paper/JN72IUBQ"},"agent_actions":{"view_html":"https://pith.science/pith/JN72IUBQH2RYBZBBYAW3FBU2BQ","download_json":"https://pith.science/pith/JN72IUBQH2RYBZBBYAW3FBU2BQ.json","view_paper":"https://pith.science/paper/JN72IUBQ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2509.06026&json=true","fetch_graph":"https://pith.science/api/pith-number/JN72IUBQH2RYBZBBYAW3FBU2BQ/graph.json","fetch_events":"https://pith.science/api/pith-number/JN72IUBQH2RYBZBBYAW3FBU2BQ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/JN72IUBQH2RYBZBBYAW3FBU2BQ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/JN72IUBQH2RYBZBBYAW3FBU2BQ/action/storage_attestation","attest_author":"https://pith.science/pith/JN72IUBQH2RYBZBBYAW3FBU2BQ/action/author_attestation","sign_citation":"https://pith.science/pith/JN72IUBQH2RYBZBBYAW3FBU2BQ/action/citation_signature","submit_replication":"https://pith.science/pith/JN72IUBQH2RYBZBBYAW3FBU2BQ/action/replication_record"}},"created_at":"2026-07-05T12:06:28.555397+00:00","updated_at":"2026-07-05T12:06:28.555397+00:00"}