{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:WTHE6N2AGVD3EY4JKFASC3VLHQ","short_pith_number":"pith:WTHE6N2A","schema_version":"1.0","canonical_sha256":"b4ce4f37403547b263895141216eab3c2cef08871a6a0ff7067e4cb9d2dd4b2a","source":{"kind":"arxiv","id":"2206.07018","version":3},"attestation_state":"computed","paper":{"title":"Turning a Curse into a Blessing: Enabling In-Distribution-Data-Free Backdoor Removal via Stabilized Model Inversion","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Jiachen T.Wang, Lingjuan Lyu, Ruoxi Jia, Si Chen, Won Park, Xun Chen, Yi Zeng, Zhuoqing Mao","submitted_at":"2022-06-14T17:32:04Z","abstract_excerpt":"Many backdoor removal techniques in machine learning models require clean in-distribution data, which may not always be available due to proprietary datasets. Model inversion techniques, often considered privacy threats, can reconstruct realistic training samples, potentially eliminating the need for in-distribution data. Prior attempts to combine backdoor removal and model inversion yielded limited results. Our work is the first to provide a thorough understanding of leveraging model inversion for effective backdoor removal by addressing key questions about reconstructed samples' properties, "},"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":"2206.07018","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2022-06-14T17:32:04Z","cross_cats_sorted":[],"title_canon_sha256":"810b4dfe513142241869173039140760ffae5638d654c20a125284bf46005d37","abstract_canon_sha256":"33114d39113f6a514bf9c80849e652243076aa4abc1e472e0f6e07604d57fce7"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T05:54:11.807926Z","signature_b64":"rJNfNvAqLHoa9uku/FxxZomzW9PFgjK3OSAsoOmYmKXSrjI1z1K00uNtUwy4Rgdyqv8TwDpV6jwp/10HtBadBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b4ce4f37403547b263895141216eab3c2cef08871a6a0ff7067e4cb9d2dd4b2a","last_reissued_at":"2026-07-05T05:54:11.807439Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T05:54:11.807439Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Turning a Curse into a Blessing: Enabling In-Distribution-Data-Free Backdoor Removal via Stabilized Model Inversion","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Jiachen T.Wang, Lingjuan Lyu, Ruoxi Jia, Si Chen, Won Park, Xun Chen, Yi Zeng, Zhuoqing Mao","submitted_at":"2022-06-14T17:32:04Z","abstract_excerpt":"Many backdoor removal techniques in machine learning models require clean in-distribution data, which may not always be available due to proprietary datasets. Model inversion techniques, often considered privacy threats, can reconstruct realistic training samples, potentially eliminating the need for in-distribution data. Prior attempts to combine backdoor removal and model inversion yielded limited results. Our work is the first to provide a thorough understanding of leveraging model inversion for effective backdoor removal by addressing key questions about reconstructed samples' properties, "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2206.07018","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/2206.07018/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":"2206.07018","created_at":"2026-07-05T05:54:11.807498+00:00"},{"alias_kind":"arxiv_version","alias_value":"2206.07018v3","created_at":"2026-07-05T05:54:11.807498+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2206.07018","created_at":"2026-07-05T05:54:11.807498+00:00"},{"alias_kind":"pith_short_12","alias_value":"WTHE6N2AGVD3","created_at":"2026-07-05T05:54:11.807498+00:00"},{"alias_kind":"pith_short_16","alias_value":"WTHE6N2AGVD3EY4J","created_at":"2026-07-05T05:54:11.807498+00:00"},{"alias_kind":"pith_short_8","alias_value":"WTHE6N2A","created_at":"2026-07-05T05:54:11.807498+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/WTHE6N2AGVD3EY4JKFASC3VLHQ","json":"https://pith.science/pith/WTHE6N2AGVD3EY4JKFASC3VLHQ.json","graph_json":"https://pith.science/api/pith-number/WTHE6N2AGVD3EY4JKFASC3VLHQ/graph.json","events_json":"https://pith.science/api/pith-number/WTHE6N2AGVD3EY4JKFASC3VLHQ/events.json","paper":"https://pith.science/paper/WTHE6N2A"},"agent_actions":{"view_html":"https://pith.science/pith/WTHE6N2AGVD3EY4JKFASC3VLHQ","download_json":"https://pith.science/pith/WTHE6N2AGVD3EY4JKFASC3VLHQ.json","view_paper":"https://pith.science/paper/WTHE6N2A","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2206.07018&json=true","fetch_graph":"https://pith.science/api/pith-number/WTHE6N2AGVD3EY4JKFASC3VLHQ/graph.json","fetch_events":"https://pith.science/api/pith-number/WTHE6N2AGVD3EY4JKFASC3VLHQ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/WTHE6N2AGVD3EY4JKFASC3VLHQ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/WTHE6N2AGVD3EY4JKFASC3VLHQ/action/storage_attestation","attest_author":"https://pith.science/pith/WTHE6N2AGVD3EY4JKFASC3VLHQ/action/author_attestation","sign_citation":"https://pith.science/pith/WTHE6N2AGVD3EY4JKFASC3VLHQ/action/citation_signature","submit_replication":"https://pith.science/pith/WTHE6N2AGVD3EY4JKFASC3VLHQ/action/replication_record"}},"created_at":"2026-07-05T05:54:11.807498+00:00","updated_at":"2026-07-05T05:54:11.807498+00:00"}