{"id":"https://openalex.org/W6946089209","doi":"https://doi.org/10.25740/kh752sm9123","title":"Identifying and Eliminating CSAM in Generative ML Training Data and Models","display_name":"Identifying and Eliminating CSAM in Generative ML Training Data and Models","publication_year":2023,"publication_date":"2023-12-20","ids":{"openalex":"https://openalex.org/W6946089209","doi":"https://doi.org/10.25740/kh752sm9123"},"language":"en","primary_location":{"id":"doi:10.25740/kh752sm9123","doi":"https://doi.org/10.25740/kh752sm9123","is_oa":true,"landing_page_url":"https://doi.org/10.25740/kh752sm9123","pdf_url":null,"source":{"id":"https://openalex.org/S7407052831","display_name":"Stanford Digital Repository","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"article-journal"},"volume":null,"type":"article","is_traditional":true,"indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.25740/kh752sm9123","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":null,"display_name":"Thiel, David","orcid":"https://orcid.org/0000-0002-0947-5921","observed_orcids":["https://orcid.org/0000-0002-0947-5921"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Thiel, David","raw_affiliation_strings":[],"raw_orcid":"https://orcid.org/0000-0002-0947-5921","affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5045528962","display_name":"Jeffrey T. Hancock","orcid":"https://orcid.org/0000-0001-5367-2677","observed_orcids":["https://orcid.org/0000-0001-5367-2677"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Hancock, Jeffrey","raw_affiliation_strings":[],"raw_orcid":"https://orcid.org/0000-0001-5367-2677","affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":0,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":2.4313,"has_fulltext":false,"cited_by_count":33,"citation_normalized_percentile":{"value":0.90551857,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":98,"max":99},"biblio":{"volume":null,"issue":null,"first_page":null,"last_page":null},"is_retracted":false,"is_paratext":false,"is_xpac":true,"primary_topic":{"id":"https://openalex.org/T12357","display_name":"Digital Media Forensic Detection","score":0.5503000020980835,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},"topics":[{"id":"https://openalex.org/T12357","display_name":"Digital Media Forensic Detection","score":0.5503000020980835,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T12262","display_name":"Hate Speech and Cyberbullying Detection","score":0.17659999430179596,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.053300000727176666,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/child-sexual-abuse-material","display_name":"child pornography","score":0.9980000257492065},{"id":"https://openalex.org/keywords/generative-ai-training-data","display_name":"generative AI training data","score":0.828000009059906},{"id":"https://openalex.org/keywords/stable-diffusion","display_name":"Stable Diffusion","score":0.9950000047683716},{"id":"https://openalex.org/keywords/data-poisoning-attacks","display_name":"data poisoning","score":0.36899998784065247},{"id":"https://openalex.org/keywords/training-data-integrity","display_name":"training data integrity","score":0.4050000011920929},{"id":"https://openalex.org/keywords/perceptual-hashing","display_name":"perceptual hashing","score":0.6570000052452087},{"id":"https://openalex.org/keywords/image-classification","display_name":"image classification","score":0.3050000071525574}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6733999848365784},{"id":"https://openalex.org/C99138194","wikidata":"https://www.wikidata.org/wiki/Q183427","display_name":"Hash function","level":2,"score":0.5896000266075134},{"id":"https://openalex.org/C2777211547","wikidata":"https://www.wikidata.org/wiki/Q17141490","display_name":"Training (meteorology)","level":2,"score":0.5514000058174133},{"id":"https://openalex.org/C39890363","wikidata":"https://www.wikidata.org/wiki/Q36108","display_name":"Generative grammar","level":2,"score":0.5307000279426575},{"id":"https://openalex.org/C98045186","wikidata":"https://www.wikidata.org/wiki/Q205663","display_name":"Process (computing)","level":2,"score":0.5303000211715698},{"id":"https://openalex.org/C167966045","wikidata":"https://www.wikidata.org/wiki/Q5532625","display_name":"Generative model","level":3,"score":0.517300009727478},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.513700008392334},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.41749998927116394},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.3698999881744385},{"id":"https://openalex.org/C2780009758","wikidata":"https://www.wikidata.org/wiki/Q6804172","display_name":"Measure (data warehouse)","level":2,"score":0.2980000078678131},{"id":"https://openalex.org/C143587482","wikidata":"https://www.wikidata.org/wiki/Q1543216","display_name":"Iterative and incremental development","level":2,"score":0.2953000068664551},{"id":"https://openalex.org/C99498987","wikidata":"https://www.wikidata.org/wiki/Q2210247","display_name":"Noise (video)","level":3,"score":0.2874000072479248},{"id":"https://openalex.org/C178489894","wikidata":"https://www.wikidata.org/wiki/Q8789","display_name":"Cryptography","level":2,"score":0.2777999937534332},{"id":"https://openalex.org/C26760741","wikidata":"https://www.wikidata.org/wiki/Q160402","display_name":"Perception","level":2,"score":0.27570000290870667},{"id":"https://openalex.org/C67186912","wikidata":"https://www.wikidata.org/wiki/Q367664","display_name":"Data modeling","level":2,"score":0.26669999957084656},{"id":"https://openalex.org/C7608002","wikidata":"https://www.wikidata.org/wiki/Q477202","display_name":"Cryptographic hash function","level":3,"score":0.2621999979019165}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.25740/kh752sm9123","doi":"https://doi.org/10.25740/kh752sm9123","is_oa":true,"landing_page_url":"https://doi.org/10.25740/kh752sm9123","pdf_url":null,"source":{"id":"https://openalex.org/S7407052831","display_name":"Stanford Digital Repository","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"article-journal"}],"best_oa_location":{"id":"doi:10.25740/kh752sm9123","doi":"https://doi.org/10.25740/kh752sm9123","is_oa":true,"landing_page_url":"https://doi.org/10.25740/kh752sm9123","pdf_url":null,"source":{"id":"https://openalex.org/S7407052831","display_name":"Stanford Digital Repository","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"article-journal"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W6964758640","https://openalex.org/W7171727041","https://openalex.org/W4417143301","https://openalex.org/W7160313311","https://openalex.org/W3203737321","https://openalex.org/W7117284073","https://openalex.org/W7162122023","https://openalex.org/W7158181690","https://openalex.org/W4413147180","https://openalex.org/W7128013653"],"abstract_inverted_index":{"Generative":[0],"Machine":[1],"Learning":[2],"models":[3,78,153,158],"have":[4,69],"been":[5],"well":[6,23,116],"documented":[7],"as":[8,22,24,115,117],"being":[9],"able":[10],"to":[11,25,33,51,61,63,140],"produce":[12,34],"explicit":[13,37],"adult":[14],"content,":[15],"including":[16],"child":[17],"sexual":[18],"abuse":[19],"material":[20],"(CSAM)":[21],"alter":[26],"benign":[27],"imagery":[28],"of":[29,47,58,77,87,106,108,143,157],"a":[30,71,85],"clothed":[31],"victim":[32],"nude":[35],"or":[36],"content.":[38],"In":[39],"this":[40,81,134,144],"study,":[41],"we":[42],"examine":[43],"the":[44,53,74,112,155],"LAION-5B":[45],"dataset\u2014parts":[46],"which":[48],"were":[49,122],"used":[50],"train":[52],"popular":[54],"Stable":[55],"Diffusion":[56],"series":[57],"models\u2014to":[59],"attempt":[60],"measure":[62],"what":[64],"degree":[65],"CSAM":[66,110],"itself":[67],"may":[68],"played":[70],"role":[72],"in":[73,111],"training":[75,113,145,149],"process":[76],"trained":[79,159],"on":[80,160],"dataset.":[82],"We":[83,128],"use":[84],"combination":[86],"PhotoDNA":[88],"perceptual":[89],"hash":[90,93],"matching,":[91,94],"cryptographic":[92],"k-nearest":[95],"neighbors":[96],"queries":[97],"and":[98,154],"ML":[99],"classifiers.":[100],"This":[101],"methodology":[102],"detected":[103],"many":[104,118],"hundreds":[105],"instances":[107],"known":[109],"set,":[114,146],"new":[119],"candidates":[120],"that":[121,138],"subsequently":[123],"verified":[124],"by":[125],"outside":[126],"parties.":[127],"also":[129],"provide":[130],"recommendations":[131],"for":[132,136],"mitigating":[133],"issue":[135],"those":[137],"need":[139],"maintain":[141],"copies":[142],"building":[147],"future":[148],"sets,":[150],"altering":[151],"existing":[152],"hosting":[156],"LAION-5B.":[161]},"counts_by_year":[{"year":2026,"cited_by_count":12},{"year":2025,"cited_by_count":14},{"year":2024,"cited_by_count":7}],"updated_date":"2026-11-07T11:25:07.637180","created_date":"2025-10-10T00:00:00"}