{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:34QDCFTSFZILBJSECEBFVC5TMD","short_pith_number":"pith:34QDCFTS","schema_version":"1.0","canonical_sha256":"df203116722e50b0a64411025a8bb360f0009aa45e9e7e3ed9d0adbdeccbcbac","source":{"kind":"arxiv","id":"2311.02373","version":2},"attestation_state":"computed","paper":{"title":"From Trojan Horses to Castle Walls: Unveiling Bilateral Data Poisoning Effects in Diffusion Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Bingquan Shen, Gaowen Liu, H. Vicky Zhao, Ramana Rao Kompella, Sijia Liu, Yuguang Yao, Zhuoshi Pan","submitted_at":"2023-11-04T11:00:31Z","abstract_excerpt":"While state-of-the-art diffusion models (DMs) excel in image generation, concerns regarding their security persist. Earlier research highlighted DMs' vulnerability to data poisoning attacks, but these studies placed stricter requirements than conventional methods like `BadNets' in image classification. This is because the art necessitates modifications to the diffusion training and sampling procedures. Unlike the prior work, we investigate whether BadNets-like data poisoning methods can directly degrade the generation by DMs. In other words, if only the training dataset is contaminated (withou"},"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":"2311.02373","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2023-11-04T11:00:31Z","cross_cats_sorted":[],"title_canon_sha256":"1c3f37874fdf4c28f5a7f8696af0d9d127ea115326edec25f32b10d1a5f7aeae","abstract_canon_sha256":"16694908af3706fb67e314274b975a5c3626e7b366e4948a042e729c8c041190"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:32:09.755510Z","signature_b64":"k4lkJ6bC34qz7Xd3xOrFT8zNqQftR7Q3xGeGKBPUIQpC/VKvb4BOxJxYjk3R+IWgjKD5B4LsloDK/b4e1JCYCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"df203116722e50b0a64411025a8bb360f0009aa45e9e7e3ed9d0adbdeccbcbac","last_reissued_at":"2026-07-05T08:32:09.754921Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:32:09.754921Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"From Trojan Horses to Castle Walls: Unveiling Bilateral Data Poisoning Effects in Diffusion Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Bingquan Shen, Gaowen Liu, H. Vicky Zhao, Ramana Rao Kompella, Sijia Liu, Yuguang Yao, Zhuoshi Pan","submitted_at":"2023-11-04T11:00:31Z","abstract_excerpt":"While state-of-the-art diffusion models (DMs) excel in image generation, concerns regarding their security persist. Earlier research highlighted DMs' vulnerability to data poisoning attacks, but these studies placed stricter requirements than conventional methods like `BadNets' in image classification. This is because the art necessitates modifications to the diffusion training and sampling procedures. Unlike the prior work, we investigate whether BadNets-like data poisoning methods can directly degrade the generation by DMs. In other words, if only the training dataset is contaminated (withou"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2311.02373","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/2311.02373/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":"2311.02373","created_at":"2026-07-05T08:32:09.754984+00:00"},{"alias_kind":"arxiv_version","alias_value":"2311.02373v2","created_at":"2026-07-05T08:32:09.754984+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2311.02373","created_at":"2026-07-05T08:32:09.754984+00:00"},{"alias_kind":"pith_short_12","alias_value":"34QDCFTSFZIL","created_at":"2026-07-05T08:32:09.754984+00:00"},{"alias_kind":"pith_short_16","alias_value":"34QDCFTSFZILBJSE","created_at":"2026-07-05T08:32:09.754984+00:00"},{"alias_kind":"pith_short_8","alias_value":"34QDCFTS","created_at":"2026-07-05T08:32:09.754984+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.09151","citing_title":"Customization under Fire: Plugin Poisoning in Text-to-Image Ecosystem","ref_index":64,"is_internal_anchor":false},{"citing_arxiv_id":"2409.18169","citing_title":"Harmful Fine-tuning Attacks and Defenses for Large Language Models: A Survey","ref_index":113,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/34QDCFTSFZILBJSECEBFVC5TMD","json":"https://pith.science/pith/34QDCFTSFZILBJSECEBFVC5TMD.json","graph_json":"https://pith.science/api/pith-number/34QDCFTSFZILBJSECEBFVC5TMD/graph.json","events_json":"https://pith.science/api/pith-number/34QDCFTSFZILBJSECEBFVC5TMD/events.json","paper":"https://pith.science/paper/34QDCFTS"},"agent_actions":{"view_html":"https://pith.science/pith/34QDCFTSFZILBJSECEBFVC5TMD","download_json":"https://pith.science/pith/34QDCFTSFZILBJSECEBFVC5TMD.json","view_paper":"https://pith.science/paper/34QDCFTS","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2311.02373&json=true","fetch_graph":"https://pith.science/api/pith-number/34QDCFTSFZILBJSECEBFVC5TMD/graph.json","fetch_events":"https://pith.science/api/pith-number/34QDCFTSFZILBJSECEBFVC5TMD/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/34QDCFTSFZILBJSECEBFVC5TMD/action/timestamp_anchor","attest_storage":"https://pith.science/pith/34QDCFTSFZILBJSECEBFVC5TMD/action/storage_attestation","attest_author":"https://pith.science/pith/34QDCFTSFZILBJSECEBFVC5TMD/action/author_attestation","sign_citation":"https://pith.science/pith/34QDCFTSFZILBJSECEBFVC5TMD/action/citation_signature","submit_replication":"https://pith.science/pith/34QDCFTSFZILBJSECEBFVC5TMD/action/replication_record"}},"created_at":"2026-07-05T08:32:09.754984+00:00","updated_at":"2026-07-05T08:32:09.754984+00:00"}