{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:Q52VDM22KTAVT6T3CGMKWG4QKE","short_pith_number":"pith:Q52VDM22","schema_version":"1.0","canonical_sha256":"877551b35a54c159fa7b1198ab1b9051251303272cbb6eac4780e0689c968e43","source":{"kind":"arxiv","id":"2204.05255","version":2},"attestation_state":"computed","paper":{"title":"Narcissus: A Practical Clean-Label Backdoor Attack with Limited Information","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CR","authors_text":"Hoang Anh Just, Lingjuan Lyu, Meikang Qiu, Minzhou Pan, Ruoxi Jia, Yi Zeng","submitted_at":"2022-04-11T16:58:04Z","abstract_excerpt":"Backdoor attacks insert malicious data into a training set so that, during inference time, it misclassifies inputs that have been patched with a backdoor trigger as the malware specified label. For backdoor attacks to bypass human inspection, it is essential that the injected data appear to be correctly labeled. The attacks with such property are often referred to as \"clean-label attacks.\" Existing clean-label backdoor attacks require knowledge of the entire training set to be effective. Obtaining such knowledge is difficult or impossible because training data are often gathered from multiple "},"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":"2204.05255","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CR","submitted_at":"2022-04-11T16:58:04Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"246f09b8ed663d12b78b2460405b8b8c7c65e5d320ba30690540ddd0033aa902","abstract_canon_sha256":"f2ec829389c4d2db10c9ebf621a7ba762a99da8bb5f93c9c2b372194bd6ffc51"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T04:15:01.906297Z","signature_b64":"4qNZBtI+uoeEC8dpRv5jODeEcX88aM2GUrAIPxS54bAGY/WuZmd41kCcqyXTwHkrFOXZb2n3pB+1uBMut7SOBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"877551b35a54c159fa7b1198ab1b9051251303272cbb6eac4780e0689c968e43","last_reissued_at":"2026-07-05T04:15:01.905906Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T04:15:01.905906Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Narcissus: A Practical Clean-Label Backdoor Attack with Limited Information","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CR","authors_text":"Hoang Anh Just, Lingjuan Lyu, Meikang Qiu, Minzhou Pan, Ruoxi Jia, Yi Zeng","submitted_at":"2022-04-11T16:58:04Z","abstract_excerpt":"Backdoor attacks insert malicious data into a training set so that, during inference time, it misclassifies inputs that have been patched with a backdoor trigger as the malware specified label. For backdoor attacks to bypass human inspection, it is essential that the injected data appear to be correctly labeled. The attacks with such property are often referred to as \"clean-label attacks.\" Existing clean-label backdoor attacks require knowledge of the entire training set to be effective. Obtaining such knowledge is difficult or impossible because training data are often gathered from multiple "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2204.05255","kind":"arxiv","version":2},"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/2204.05255/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":"2204.05255","created_at":"2026-07-05T04:15:01.905960+00:00"},{"alias_kind":"arxiv_version","alias_value":"2204.05255v2","created_at":"2026-07-05T04:15:01.905960+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2204.05255","created_at":"2026-07-05T04:15:01.905960+00:00"},{"alias_kind":"pith_short_12","alias_value":"Q52VDM22KTAV","created_at":"2026-07-05T04:15:01.905960+00:00"},{"alias_kind":"pith_short_16","alias_value":"Q52VDM22KTAVT6T3","created_at":"2026-07-05T04:15:01.905960+00:00"},{"alias_kind":"pith_short_8","alias_value":"Q52VDM22","created_at":"2026-07-05T04:15:01.905960+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2411.10673","citing_title":"How to Defend Against Large-scale Model Poisoning Attacks in Federated Learning: A Vertical Solution","ref_index":2022,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/Q52VDM22KTAVT6T3CGMKWG4QKE","json":"https://pith.science/pith/Q52VDM22KTAVT6T3CGMKWG4QKE.json","graph_json":"https://pith.science/api/pith-number/Q52VDM22KTAVT6T3CGMKWG4QKE/graph.json","events_json":"https://pith.science/api/pith-number/Q52VDM22KTAVT6T3CGMKWG4QKE/events.json","paper":"https://pith.science/paper/Q52VDM22"},"agent_actions":{"view_html":"https://pith.science/pith/Q52VDM22KTAVT6T3CGMKWG4QKE","download_json":"https://pith.science/pith/Q52VDM22KTAVT6T3CGMKWG4QKE.json","view_paper":"https://pith.science/paper/Q52VDM22","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2204.05255&json=true","fetch_graph":"https://pith.science/api/pith-number/Q52VDM22KTAVT6T3CGMKWG4QKE/graph.json","fetch_events":"https://pith.science/api/pith-number/Q52VDM22KTAVT6T3CGMKWG4QKE/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/Q52VDM22KTAVT6T3CGMKWG4QKE/action/timestamp_anchor","attest_storage":"https://pith.science/pith/Q52VDM22KTAVT6T3CGMKWG4QKE/action/storage_attestation","attest_author":"https://pith.science/pith/Q52VDM22KTAVT6T3CGMKWG4QKE/action/author_attestation","sign_citation":"https://pith.science/pith/Q52VDM22KTAVT6T3CGMKWG4QKE/action/citation_signature","submit_replication":"https://pith.science/pith/Q52VDM22KTAVT6T3CGMKWG4QKE/action/replication_record"}},"created_at":"2026-07-05T04:15:01.905960+00:00","updated_at":"2026-07-05T04:15:01.905960+00:00"}