{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:ML2Z4XABC3N3PSXJTJVYLTL4I2","short_pith_number":"pith:ML2Z4XAB","schema_version":"1.0","canonical_sha256":"62f59e5c0116dbb7cae99a6b85cd7c468112192cf3a5410d05737fc5f1b38e24","source":{"kind":"arxiv","id":"2409.05294","version":1},"attestation_state":"computed","paper":{"title":"TERD: A Unified Framework for Safeguarding Diffusion Models Against Backdoors","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CR","authors_text":"Ang Li, Hui Huang, Mingjie Li, Yichuan Mo, Yisen Wang","submitted_at":"2024-09-09T03:02:16Z","abstract_excerpt":"Diffusion models have achieved notable success in image generation, but they remain highly vulnerable to backdoor attacks, which compromise their integrity by producing specific undesirable outputs when presented with a pre-defined trigger. In this paper, we investigate how to protect diffusion models from this dangerous threat. Specifically, we propose TERD, a backdoor defense framework that builds unified modeling for current attacks, which enables us to derive an accessible reversed loss. A trigger reversion strategy is further employed: an initial approximation of the trigger through noise"},"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":"2409.05294","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CR","submitted_at":"2024-09-09T03:02:16Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"2e07fd9142b79063a645a037e68d12f75ce035bef29e7d2fa495b3769e9a28ac","abstract_canon_sha256":"b14015386d20b3dda4b58d60753ae6fac93693be7a08ac2c2a4b410b9164b708"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:04:47.277526Z","signature_b64":"Ky9ZAxH3wQt9UuSbuLiE6MSQp8t4HqpXRsm/ZAmXNjQ3vu0re619NwfcFA3XWirOCoh5MIeRZ8D3xg4iphaaAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"62f59e5c0116dbb7cae99a6b85cd7c468112192cf3a5410d05737fc5f1b38e24","last_reissued_at":"2026-07-05T09:04:47.277137Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:04:47.277137Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"TERD: A Unified Framework for Safeguarding Diffusion Models Against Backdoors","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CR","authors_text":"Ang Li, Hui Huang, Mingjie Li, Yichuan Mo, Yisen Wang","submitted_at":"2024-09-09T03:02:16Z","abstract_excerpt":"Diffusion models have achieved notable success in image generation, but they remain highly vulnerable to backdoor attacks, which compromise their integrity by producing specific undesirable outputs when presented with a pre-defined trigger. In this paper, we investigate how to protect diffusion models from this dangerous threat. Specifically, we propose TERD, a backdoor defense framework that builds unified modeling for current attacks, which enables us to derive an accessible reversed loss. A trigger reversion strategy is further employed: an initial approximation of the trigger through noise"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2409.05294","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/2409.05294/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":"2409.05294","created_at":"2026-07-05T09:04:47.277189+00:00"},{"alias_kind":"arxiv_version","alias_value":"2409.05294v1","created_at":"2026-07-05T09:04:47.277189+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2409.05294","created_at":"2026-07-05T09:04:47.277189+00:00"},{"alias_kind":"pith_short_12","alias_value":"ML2Z4XABC3N3","created_at":"2026-07-05T09:04:47.277189+00:00"},{"alias_kind":"pith_short_16","alias_value":"ML2Z4XABC3N3PSXJ","created_at":"2026-07-05T09:04:47.277189+00:00"},{"alias_kind":"pith_short_8","alias_value":"ML2Z4XAB","created_at":"2026-07-05T09:04:47.277189+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.23362","citing_title":"TooBad: Backdoor Diffusion Models with Ultra-Low Poison Rate and Imperceptible Trigger","ref_index":29,"is_internal_anchor":false},{"citing_arxiv_id":"2605.08280","citing_title":"Beyond the False Trade-off: Adaptive EWC for Stealthy and Generalizable T2I Backdoors","ref_index":15,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/ML2Z4XABC3N3PSXJTJVYLTL4I2","json":"https://pith.science/pith/ML2Z4XABC3N3PSXJTJVYLTL4I2.json","graph_json":"https://pith.science/api/pith-number/ML2Z4XABC3N3PSXJTJVYLTL4I2/graph.json","events_json":"https://pith.science/api/pith-number/ML2Z4XABC3N3PSXJTJVYLTL4I2/events.json","paper":"https://pith.science/paper/ML2Z4XAB"},"agent_actions":{"view_html":"https://pith.science/pith/ML2Z4XABC3N3PSXJTJVYLTL4I2","download_json":"https://pith.science/pith/ML2Z4XABC3N3PSXJTJVYLTL4I2.json","view_paper":"https://pith.science/paper/ML2Z4XAB","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2409.05294&json=true","fetch_graph":"https://pith.science/api/pith-number/ML2Z4XABC3N3PSXJTJVYLTL4I2/graph.json","fetch_events":"https://pith.science/api/pith-number/ML2Z4XABC3N3PSXJTJVYLTL4I2/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/ML2Z4XABC3N3PSXJTJVYLTL4I2/action/timestamp_anchor","attest_storage":"https://pith.science/pith/ML2Z4XABC3N3PSXJTJVYLTL4I2/action/storage_attestation","attest_author":"https://pith.science/pith/ML2Z4XABC3N3PSXJTJVYLTL4I2/action/author_attestation","sign_citation":"https://pith.science/pith/ML2Z4XABC3N3PSXJTJVYLTL4I2/action/citation_signature","submit_replication":"https://pith.science/pith/ML2Z4XABC3N3PSXJTJVYLTL4I2/action/replication_record"}},"created_at":"2026-07-05T09:04:47.277189+00:00","updated_at":"2026-07-05T09:04:47.277189+00:00"}