{"id":"https://openalex.org/W7233774542","doi":"https://doi.org/10.1109/tdsc.2026.3754070","title":"DeepSanitizer: Data-free Backdoor Removal in Deep Neural Networks","display_name":"DeepSanitizer: Data-free Backdoor Removal in Deep Neural Networks","alternate_titles":[],"publication_year":2026,"publication_date":"2026-01-01","ids":{"openalex":"https://openalex.org/W7233774542","doi":"https://doi.org/10.1109/tdsc.2026.3754070"},"language":"en","primary_location":{"id":"doi:10.1109/tdsc.2026.3754070","doi":"https://doi.org/10.1109/tdsc.2026.3754070","is_oa":false,"landing_page_url":"https://doi.org/10.1109/tdsc.2026.3754070","pdf_url":null,"source":{"id":"https://openalex.org/S133795288","display_name":"IEEE Transactions on Dependable and Secure Computing","issn_l":"1545-5971","issn":["1545-5971","1941-0018","2160-9209"],"is_oa":false,"is_in_doaj":false,"is_core":true,"listed_in":["anvur-scientific","ccf-a","cwts-core","dongbi-a","jufo-3","norway-2","poland-140","russia-white-list-1","vabb-shw"],"host_organization":"https://openalex.org/P4310320439","host_organization_name":"IEEE Computer Society","host_organization_lineage":["https://openalex.org/P4310320439","https://openalex.org/P4310319808"],"host_organization_lineage_names":["IEEE Computer Society","Institute of Electrical and Electronics Engineers"],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"IEEE Transactions on Dependable and Secure Computing","raw_type":"journal-article"},"parent_works":[],"lineage":["https://openalex.org/W7233774542"],"child_works_count":0,"type":"article","is_traditional":true,"indexed_in":["crossref"],"open_access":{"is_oa":false,"oa_status":"closed","oa_url":null,"any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5021807506","display_name":"Yiming Wu","orcid":"https://orcid.org/0000-0003-4788-1485","observed_orcids":["https://orcid.org/0000-0003-4788-1485","https://orcid.org/0009-0005-7045-9233"]},"institutions":[{"id":"https://openalex.org/I55712492","display_name":"Zhejiang University of Technology","ror":"https://ror.org/02djqfd08","country_code":"CN","type":"education","lineage":["https://openalex.org/I55712492"]}],"countries":["CN"],"is_corresponding":false,"contributor_roles":[],"raw_author_name":"Yiming Wu","raw_affiliation_strings":["Zhejiang University of Technology, Hangzhou, China"],"raw_orcid":"https://orcid.org/0000-0003-4788-1485","affiliations":[{"raw_affiliation_string":"Zhejiang University of Technology, Hangzhou, China","institution_ids":["https://openalex.org/I55712492"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5102674783","display_name":"Nanjie Yao","orcid":"https://orcid.org/0009-0007-4930-7758","observed_orcids":["https://orcid.org/0009-0007-4930-7758"]},"institutions":[{"id":"https://openalex.org/I55712492","display_name":"Zhejiang University of Technology","ror":"https://ror.org/02djqfd08","country_code":"CN","type":"education","lineage":["https://openalex.org/I55712492"]}],"countries":["CN"],"is_corresponding":false,"contributor_roles":[],"raw_author_name":"Nanjie Yao","raw_affiliation_strings":["Zhejiang University of Technology, Hangzhou, China"],"raw_orcid":"https://orcid.org/0009-0007-4930-7758","affiliations":[{"raw_affiliation_string":"Zhejiang University of Technology, Hangzhou, China","institution_ids":["https://openalex.org/I55712492"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5103201254","display_name":"Chong Long Fu","orcid":"https://orcid.org/0000-0001-6465-2305","observed_orcids":["https://orcid.org/0000-0001-6465-2305"]},"institutions":[{"id":"https://openalex.org/I76130692","display_name":"Zhejiang University","ror":"https://ror.org/00a2xv884","country_code":"CN","type":"education","lineage":["https://openalex.org/I76130692"]}],"countries":["CN"],"is_corresponding":false,"contributor_roles":[],"raw_author_name":"Chong Fu","raw_affiliation_strings":["Zhejiang University, Hangzhou, China"],"raw_orcid":"https://orcid.org/0000-0001-6465-2305","affiliations":[{"raw_affiliation_string":"Zhejiang University, Hangzhou, China","institution_ids":["https://openalex.org/I76130692"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5016502133","display_name":"Chenghui Shi","orcid":"https://orcid.org/0000-0001-8799-9045","observed_orcids":["https://orcid.org/0000-0001-8799-9045"]},"institutions":[{"id":"https://openalex.org/I76130692","display_name":"Zhejiang University","ror":"https://ror.org/00a2xv884","country_code":"CN","type":"education","lineage":["https://openalex.org/I76130692"]}],"countries":["CN"],"is_corresponding":false,"contributor_roles":[],"raw_author_name":"Chenghui Shi","raw_affiliation_strings":["Zhejiang University, Hangzhou, China"],"raw_orcid":"https://orcid.org/0000-0001-8799-9045","affiliations":[{"raw_affiliation_string":"Zhejiang University, Hangzhou, China","institution_ids":["https://openalex.org/I76130692"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5102731573","display_name":"Tiejun Wu","orcid":"https://orcid.org/0000-0001-7748-964X","observed_orcids":["https://orcid.org/0000-0001-7748-964X"]},"institutions":[],"countries":[],"is_corresponding":false,"contributor_roles":[],"raw_author_name":"Tiejun Wu","raw_affiliation_strings":["NSFOCUS Technologies Group Co., Ltd, Beijing, China"],"raw_orcid":"https://orcid.org/0000-0001-7748-964X","affiliations":[{"raw_affiliation_string":"NSFOCUS Technologies Group Co., Ltd, Beijing, China","institution_ids":[]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5058611515","display_name":"Shouling Ji","orcid":"https://orcid.org/0000-0003-4268-372X","observed_orcids":["https://orcid.org/0000-0003-4268-372X"]},"institutions":[{"id":"https://openalex.org/I76130692","display_name":"Zhejiang University","ror":"https://ror.org/00a2xv884","country_code":"CN","type":"education","lineage":["https://openalex.org/I76130692"]}],"countries":["CN"],"is_corresponding":false,"contributor_roles":[],"raw_author_name":"Shouling Ji","raw_affiliation_strings":["Zhejiang University, Hangzhou, China"],"raw_orcid":"https://orcid.org/0000-0003-4268-372X","affiliations":[{"raw_affiliation_string":"Zhejiang University, Hangzhou, China","institution_ids":["https://openalex.org/I76130692"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":2,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":{"value":2645,"currency":"USD","value_usd":2645},"apc_paid":null,"fwci":0.0,"has_fulltext":false,"cited_by_count":0,"reviewed_by_count":0,"citation_normalized_percentile":{"value":0.0,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"1","last_page":"18"},"is_retracted":false,"is_pulled":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.9994999766349792,"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"}},"topics":[{"id":"https://openalex.org/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.9994999766349792,"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/T11424","display_name":"Security and Verification in Computing","score":0.00019999999494757503,"subfield":{"id":"https://openalex.org/subfields/1708","display_name":"Hardware and Architecture"},"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/T12122","display_name":"Physical Unclonable Functions (PUFs) and Hardware Security","score":9.999999747378752e-05,"subfield":{"id":"https://openalex.org/subfields/1708","display_name":"Hardware and Architecture"},"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/backdoor-removal","display_name":"backdoor removal","score":0.8360000252723694},{"id":"https://openalex.org/keywords/backdoor-attacks","display_name":"backdoor attacks","score":0.9599999785423279},{"id":"https://openalex.org/keywords/deep-neural-network","display_name":"deep neural network","score":0.7260000109672546}],"concepts":[{"id":"https://openalex.org/C2781045450","wikidata":"https://www.wikidata.org/wiki/Q254569","display_name":"Backdoor","level":2,"score":0.9979000091552734},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8003000020980835},{"id":"https://openalex.org/C95623464","wikidata":"https://www.wikidata.org/wiki/Q1096149","display_name":"Classifier (UML)","level":2,"score":0.6161999702453613},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5892000198364258},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.5206000208854675},{"id":"https://openalex.org/C117978034","wikidata":"https://www.wikidata.org/wiki/Q5422192","display_name":"Extractor","level":2,"score":0.4837999939918518},{"id":"https://openalex.org/C174333608","wikidata":"https://www.wikidata.org/wiki/Q19635","display_name":"Trojan","level":2,"score":0.4767000079154968},{"id":"https://openalex.org/C108583219","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep learning","level":2,"score":0.46799999475479126},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.4652000069618225},{"id":"https://openalex.org/C2776401178","wikidata":"https://www.wikidata.org/wiki/Q12050496","display_name":"Feature (linguistics)","level":2,"score":0.45829999446868896},{"id":"https://openalex.org/C2984842247","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep neural networks","level":3,"score":0.3993000090122223},{"id":"https://openalex.org/C52622490","wikidata":"https://www.wikidata.org/wiki/Q1026626","display_name":"Feature extraction","level":2,"score":0.37139999866485596},{"id":"https://openalex.org/C95713431","wikidata":"https://www.wikidata.org/wiki/Q631425","display_name":"Vulnerability (computing)","level":2,"score":0.3303000032901764},{"id":"https://openalex.org/C2779797433","wikidata":"https://www.wikidata.org/wiki/Q632959","display_name":"Blacklisting","level":2,"score":0.3287999927997589},{"id":"https://openalex.org/C63479239","wikidata":"https://www.wikidata.org/wiki/Q7353546","display_name":"Robustness (evolution)","level":3,"score":0.320499986410141},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.28679999709129333},{"id":"https://openalex.org/C67186912","wikidata":"https://www.wikidata.org/wiki/Q367664","display_name":"Data modeling","level":2,"score":0.27399998903274536},{"id":"https://openalex.org/C153180895","wikidata":"https://www.wikidata.org/wiki/Q7148389","display_name":"Pattern recognition (psychology)","level":2,"score":0.27379998564720154},{"id":"https://openalex.org/C59404180","wikidata":"https://www.wikidata.org/wiki/Q17013334","display_name":"Feature learning","level":2,"score":0.25760000944137573}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/tdsc.2026.3754070","doi":"https://doi.org/10.1109/tdsc.2026.3754070","is_oa":false,"landing_page_url":"https://doi.org/10.1109/tdsc.2026.3754070","pdf_url":null,"source":{"id":"https://openalex.org/S133795288","display_name":"IEEE Transactions on Dependable and Secure Computing","issn_l":"1545-5971","issn":["1545-5971","1941-0018","2160-9209"],"is_oa":false,"is_in_doaj":false,"is_core":true,"listed_in":["anvur-scientific","ccf-a","cwts-core","dongbi-a","jufo-3","norway-2","poland-140","russia-white-list-1","vabb-shw"],"host_organization":"https://openalex.org/P4310320439","host_organization_name":"IEEE Computer Society","host_organization_lineage":["https://openalex.org/P4310320439","https://openalex.org/P4310319808"],"host_organization_lineage_names":["IEEE Computer Society","Institute of Electrical and Electronics Engineers"],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"IEEE Transactions on Dependable and Secure Computing","raw_type":"journal-article"}],"best_oa_location":null,"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W4383221550","https://openalex.org/W7165177359","https://openalex.org/W4413463928","https://openalex.org/W7203526467","https://openalex.org/W7162260842","https://openalex.org/W4402916397","https://openalex.org/W3121478722","https://openalex.org/W4409361699","https://openalex.org/W4322760095","https://openalex.org/W4416016654"],"abstract_inverted_index":{"Deep":[0],"Neural":[1],"Networks":[2],"(DNNs)":[3],"are":[4],"widely":[5],"utilized":[6],"in":[7,79],"various":[8],"Artificial":[9],"Intelligence":[10],"(AI)":[11],"systems,":[12],"such":[13,83],"as":[14,33,84,207],"face":[15],"recognition,":[16],"speech":[17],"assistant,":[18],"and":[19,187,239],"autonomous":[20],"driving.":[21],"However,":[22,73],"recent":[23],"studies":[24],"reveal":[25],"DNN's":[26],"vulnerability":[27],"to":[28,53,64,123,133,149,168,203,216,250],"backdoor":[29,58,112,124,161,241],"attacks,":[30],"also":[31],"known":[32,130],"trojan":[34],"attacks.":[35],"A":[36],"backdoored":[37],"DNN":[38],"makes":[39],"wrong":[40],"predictions":[41],"when":[42,85],"the":[43,46,71,120,151,157,164,184,188,208,213,218,232,247,256],"input":[44],"contains":[45],"trigger,":[47],"which":[48,104],"poses":[49],"significant":[50],"security":[51],"threats":[52],"DNN-based":[54],"AI":[55],"systems.":[56],"Existing":[57],"removal":[59],"methods":[60],"often":[61],"require":[62],"access":[63,167],"clean":[65,170],"validation":[66],"samples":[67,69],"or":[68,92,136,171],"with":[70,87],"trigger.":[72],"this":[74,98],"may":[75],"not":[76],"be":[77],"feasible":[78],"some":[80],"real-world":[81],"scenarios,":[82],"dealing":[86],"models":[88,140],"from":[89],"online":[90],"platforms":[91],"federated":[93],"learning":[94],"environments.":[95],"To":[96],"bridge":[97],"critical":[99],"gap,":[100],"we":[101],"propose":[102],"DeepSanitizer,":[103],"introduces":[105],"a":[106,142,178,199],"new":[107],"algorithmic":[108],"formulation":[109],"for":[110,166,211],"data-free":[111,238],"mitigation.":[113],"Unlike":[114],"prior":[115],"detection-focused":[116],"workFreeEagle,":[117],"DeepSanitizer":[118,144,245],"shifts":[119],"paradigm":[121],"directly":[122],"removal.":[125],"Rather":[126],"than":[127,255],"merely":[128],"applying":[129],"reverseengineering":[131],"techniques":[132],"reconstruct":[134],"triggers":[135],"integrating":[137],"existing":[138,257],"pruning":[139],"into":[141,181],"pipeline,":[143],"synthesizes":[145],"benign":[146],"dummy":[147,209],"representations":[148,206],"fine-tune":[150],"compromised":[152,179],"model":[153,180,189,214,233],"classifier.":[154],"This":[155],"enables":[156],"mitigation":[158,242],"of":[159],"diverse":[160],"attacks":[162],"without":[163],"need":[165],"either":[169],"poisoned":[172],"data.":[173],"Specifically,":[174],"our":[175,224],"method":[176],"splits":[177],"two":[182],"components:":[183],"feature":[185,205],"extractor":[186],"classifier":[190,215],"via":[191],"an":[192],"introduced":[193],"inspection":[194],"layer.":[195],"Then":[196],"it":[197],"utilizes":[198],"novel":[200],"gradient-based":[201],"algorithm":[202],"generate":[204],"data":[210],"fine-tuning":[212],"remove":[217],"backdoor.":[219],"Extensive":[220],"experiments":[221],"demonstrate":[222],"that":[223],"approach":[225],"effectively":[226],"removes":[227],"planted":[228],"backdoors":[229],"while":[230],"maintaining":[231],"performance,":[234],"significantly":[235],"outperforming":[236],"state-of-the-art":[237],"non-data-free":[240],"techniques.":[243],"Finally,":[244],"reduces":[246],"defense":[248],"time":[249],"just":[251],"minutes,":[252],"greatly":[253],"faster":[254],"methods.":[258]},"cited_by_api_url":"https://api.openalex.org/works?filter=cites:W7233774542","counts_by_year":[],"updated_date":"2026-11-20T13:11:24.684918","created_date":"2026-11-17T00:00:00"}
