{"id":"https://openalex.org/W4412748185","doi":"https://doi.org/10.1109/tifs.2025.3594037","title":"Semantic and Precise Trigger Inversion: Detecting Backdoored Language Models","display_name":"Semantic and Precise Trigger Inversion: Detecting Backdoored Language Models","publication_year":2025,"publication_date":"2025-01-01","ids":{"openalex":"https://openalex.org/W4412748185","doi":"https://doi.org/10.1109/tifs.2025.3594037"},"language":"en","primary_location":{"id":"doi:10.1109/tifs.2025.3594037","is_oa":false,"landing_page_url":"https://doi.org/10.1109/tifs.2025.3594037","pdf_url":null,"source":{"id":"https://openalex.org/S61310614","display_name":"IEEE Transactions on Information Forensics and Security","issn_l":"1556-6013","issn":["1556-6013","1556-6021"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310319808","host_organization_name":"Institute of Electrical and Electronics Engineers","host_organization_lineage":["https://openalex.org/P4310319808"],"host_organization_lineage_names":["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 Information Forensics and Security","raw_type":"journal-article"},"type":"article","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/A5011130731","display_name":"Chunlong Xie","orcid":null},"institutions":[{"id":"https://openalex.org/I158842170","display_name":"Chongqing University","ror":"https://ror.org/023rhb549","country_code":"CN","type":"education","lineage":["https://openalex.org/I158842170"]}],"countries":["CN"],"is_corresponding":true,"raw_author_name":"Chunlong Xie","raw_affiliation_strings":["College of Computer Science, Chongqing University, Chongqing, China"],"raw_orcid":"https://orcid.org/0009-0000-3282-8062","affiliations":[{"raw_affiliation_string":"College of Computer Science, Chongqing University, Chongqing, China","institution_ids":["https://openalex.org/I158842170"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5016904590","display_name":"Jialing He","orcid":"https://orcid.org/0000-0002-8643-0647"},"institutions":[{"id":"https://openalex.org/I158842170","display_name":"Chongqing University","ror":"https://ror.org/023rhb549","country_code":"CN","type":"education","lineage":["https://openalex.org/I158842170"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Jialing He","raw_affiliation_strings":["College of Computer Science, Chongqing University, Chongqing, China"],"raw_orcid":"https://orcid.org/0000-0002-8643-0647","affiliations":[{"raw_affiliation_string":"College of Computer Science, Chongqing University, Chongqing, China","institution_ids":["https://openalex.org/I158842170"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5043576983","display_name":"Ying Yang","orcid":"https://orcid.org/0000-0002-0617-455X"},"institutions":[{"id":"https://openalex.org/I115228651","display_name":"Agency for Science, Technology and Research","ror":"https://ror.org/036wvzt09","country_code":"SG","type":"government","lineage":["https://openalex.org/I115228651"]},{"id":"https://openalex.org/I3004594783","display_name":"Institute of High Performance Computing","ror":"https://ror.org/02n0ejh50","country_code":"SG","type":"facility","lineage":["https://openalex.org/I115228651","https://openalex.org/I3004594783","https://openalex.org/I91275662"]}],"countries":["SG"],"is_corresponding":false,"raw_author_name":"Ying Yang","raw_affiliation_strings":["Institute of High Performance Computing (IHPC) and the Centre for Frontier AI Research (CFAR), Agency for Science, Technology and Research (A*STAR), Fusionopolis, Singapore","Institute of High Performance Computing (IHPC) and Centre for Frontier AI Research (CFAR), Agency for Science, Technology and Research (A*STAR), Singapore"],"raw_orcid":"https://orcid.org/0000-0002-0617-455X","affiliations":[{"raw_affiliation_string":"Institute of High Performance Computing (IHPC) and the Centre for Frontier AI Research (CFAR), Agency for Science, Technology and Research (A*STAR), Fusionopolis, Singapore","institution_ids":["https://openalex.org/I3004594783"]},{"raw_affiliation_string":"Institute of High Performance Computing (IHPC) and Centre for Frontier AI Research (CFAR), Agency for Science, Technology and Research (A*STAR), Singapore","institution_ids":["https://openalex.org/I3004594783","https://openalex.org/I115228651"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5073264981","display_name":"Shangwei Guo","orcid":"https://orcid.org/0000-0002-6443-5308"},"institutions":[{"id":"https://openalex.org/I158842170","display_name":"Chongqing University","ror":"https://ror.org/023rhb549","country_code":"CN","type":"education","lineage":["https://openalex.org/I158842170"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Shangwei Guo","raw_affiliation_strings":["College of Computer Science, Chongqing University, Chongqing, China"],"raw_orcid":"https://orcid.org/0000-0002-6443-5308","affiliations":[{"raw_affiliation_string":"College of Computer Science, Chongqing University, Chongqing, China","institution_ids":["https://openalex.org/I158842170"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101591101","display_name":"Tianwei Zhang","orcid":"https://orcid.org/0000-0001-6595-6650"},"institutions":[{"id":"https://openalex.org/I172675005","display_name":"Nanyang Technological University","ror":"https://ror.org/02e7b5302","country_code":"SG","type":"education","lineage":["https://openalex.org/I172675005"]}],"countries":["SG"],"is_corresponding":false,"raw_author_name":"Tianwei Zhang","raw_affiliation_strings":["College of Computing and Data Science, Nanyang Technological University, Jurong West, Singapore","College of Computing and Data Science, Nanyang Technological University, Singapore"],"raw_orcid":"https://orcid.org/0000-0001-6595-6650","affiliations":[{"raw_affiliation_string":"College of Computing and Data Science, Nanyang Technological University, Jurong West, Singapore","institution_ids":["https://openalex.org/I172675005"]},{"raw_affiliation_string":"College of Computing and Data Science, Nanyang Technological University, Singapore","institution_ids":["https://openalex.org/I172675005"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5079461118","display_name":"Tao Xiang","orcid":"https://orcid.org/0000-0002-9439-4623"},"institutions":[{"id":"https://openalex.org/I158842170","display_name":"Chongqing University","ror":"https://ror.org/023rhb549","country_code":"CN","type":"education","lineage":["https://openalex.org/I158842170"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Tao Xiang","raw_affiliation_strings":["College of Computer Science, Chongqing University, Chongqing, China"],"raw_orcid":"https://orcid.org/0000-0002-9439-4623","affiliations":[{"raw_affiliation_string":"College of Computer Science, Chongqing University, Chongqing, China","institution_ids":["https://openalex.org/I158842170"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":6,"corresponding_author_ids":["https://openalex.org/A5011130731"],"corresponding_institution_ids":["https://openalex.org/I158842170"],"apc_list":null,"apc_paid":null,"fwci":2.1733,"has_fulltext":false,"cited_by_count":1,"citation_normalized_percentile":{"value":0.89557444,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":96,"max":98},"biblio":{"volume":"20","issue":null,"first_page":"8096","last_page":"8108"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10181","display_name":"Natural Language Processing Techniques","score":0.932200014591217,"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/T10181","display_name":"Natural Language Processing Techniques","score":0.932200014591217,"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/T10028","display_name":"Topic Modeling","score":0.9175000190734863,"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/computer-science","display_name":"Computer science","score":0.7967333197593689},{"id":"https://openalex.org/keywords/inversion","display_name":"Inversion (geology)","score":0.48009851574897766},{"id":"https://openalex.org/keywords/natural-language-processing","display_name":"Natural language processing","score":0.4452574551105499},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.4044378399848938},{"id":"https://openalex.org/keywords/geology","display_name":"Geology","score":0.15860632061958313}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7967333197593689},{"id":"https://openalex.org/C1893757","wikidata":"https://www.wikidata.org/wiki/Q3653001","display_name":"Inversion (geology)","level":3,"score":0.48009851574897766},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.4452574551105499},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4044378399848938},{"id":"https://openalex.org/C127313418","wikidata":"https://www.wikidata.org/wiki/Q1069","display_name":"Geology","level":0,"score":0.15860632061958313},{"id":"https://openalex.org/C151730666","wikidata":"https://www.wikidata.org/wiki/Q7205","display_name":"Paleontology","level":1,"score":0.0},{"id":"https://openalex.org/C109007969","wikidata":"https://www.wikidata.org/wiki/Q749565","display_name":"Structural basin","level":2,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/tifs.2025.3594037","is_oa":false,"landing_page_url":"https://doi.org/10.1109/tifs.2025.3594037","pdf_url":null,"source":{"id":"https://openalex.org/S61310614","display_name":"IEEE Transactions on Information Forensics and Security","issn_l":"1556-6013","issn":["1556-6013","1556-6021"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310319808","host_organization_name":"Institute of Electrical and Electronics Engineers","host_organization_lineage":["https://openalex.org/P4310319808"],"host_organization_lineage_names":["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 Information Forensics and Security","raw_type":"journal-article"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[{"id":"https://openalex.org/G1546930127","display_name":null,"funder_award_id":"U20A20176","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"},{"id":"https://openalex.org/G217923984","display_name":null,"funder_award_id":"CQJT-CZKJ2024-03","funder_id":"https://openalex.org/F4320330209","funder_display_name":"Chongqing Municipal Development and Reform Commission"},{"id":"https://openalex.org/G3550020321","display_name":null,"funder_award_id":"U21A20463","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"},{"id":"https://openalex.org/G8711352089","display_name":null,"funder_award_id":"62472057","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"}],"funders":[{"id":"https://openalex.org/F4320321001","display_name":"National Natural Science Foundation of China","ror":"https://ror.org/01h0zpd94"},{"id":"https://openalex.org/F4320330209","display_name":"Chongqing Municipal Development and Reform Commission","ror":null}],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":45,"referenced_works":["https://openalex.org/W2088911157","https://openalex.org/W2144578941","https://openalex.org/W2753783305","https://openalex.org/W2934843808","https://openalex.org/W2963126845","https://openalex.org/W2963748441","https://openalex.org/W2965373594","https://openalex.org/W2971196067","https://openalex.org/W2973217491","https://openalex.org/W2982756474","https://openalex.org/W3034457371","https://openalex.org/W3035367371","https://openalex.org/W3042368254","https://openalex.org/W3098371839","https://openalex.org/W3100727892","https://openalex.org/W3109409894","https://openalex.org/W3128663834","https://openalex.org/W3158360872","https://openalex.org/W3171523434","https://openalex.org/W3173784240","https://openalex.org/W3175052694","https://openalex.org/W3176270593","https://openalex.org/W3195374612","https://openalex.org/W3205696278","https://openalex.org/W3207360435","https://openalex.org/W4288057740","https://openalex.org/W4385644179","https://openalex.org/W4392909875","https://openalex.org/W4396982204","https://openalex.org/W6631190155","https://openalex.org/W6676984168","https://openalex.org/W6729448088","https://openalex.org/W6743581629","https://openalex.org/W6755207826","https://openalex.org/W6761205521","https://openalex.org/W6768817161","https://openalex.org/W6768851824","https://openalex.org/W6771917389","https://openalex.org/W6779690972","https://openalex.org/W6788272337","https://openalex.org/W6791576469","https://openalex.org/W6801759574","https://openalex.org/W6810332148","https://openalex.org/W6870466612","https://openalex.org/W6946857974"],"related_works":["https://openalex.org/W4391375266","https://openalex.org/W2899084033","https://openalex.org/W2748952813","https://openalex.org/W2390279801","https://openalex.org/W4391913857","https://openalex.org/W2358668433","https://openalex.org/W4396701345","https://openalex.org/W2376932109","https://openalex.org/W2001405890","https://openalex.org/W3204019825"],"abstract_inverted_index":{"Backdoor":[0],"attacks":[1],"pose":[2],"a":[3,82,95,119,204],"serious":[4],"security":[5,201],"threat":[6],"to":[7,15,29,49,58,64,72],"Natural":[8],"Language":[9],"Processing":[10],"(NLP)":[11],"models,":[12],"allowing":[13],"adversaries":[14],"manipulate":[16],"model":[17],"outputs":[18],"through":[19],"hidden":[20],"triggers.":[21,45,135],"Although":[22],"backdoor":[23,61,179,210],"detection":[24,87,180],"methods":[25,47],"have":[26],"been":[27],"developed":[28],"address":[30,90],"this":[31],"issue,":[32],"existing":[33],"approaches":[34,176],"based":[35],"on":[36,66],"trigger":[37,97,125,183],"inversion":[38,98,110,117,131,184],"are":[39],"effective":[40,130,190],"only":[41],"for":[42,151,208],"simple,":[43],"visible":[44],"These":[46],"struggle":[48],"handle":[50],"semantically":[51,133],"enhanced,":[52],"invisible":[53],"triggers":[54],"and":[55,111,169,182],"often":[56],"fail":[57],"provide":[59],"accurate":[60],"determinations":[62],"due":[63],"reliance":[65],"unreliable":[67],"heuristics,":[68],"making":[69],"it":[70,172],"difficult":[71],"reliably":[73],"distinguish":[74],"backdoored":[75,153],"models":[76,154,192],"from":[77,155],"benign":[78,156],"ones.":[79,157],"This":[80],"presents":[81],"critical":[83],"gap":[84],"in":[85,177,216],"current":[86],"techniques.":[88],"To":[89],"these":[91],"challenges,":[92],"we":[93],"propose":[94],"novel":[96],"<italic":[99,160],"xmlns:mml=\"http://www.w3.org/1998/Math/MathML\"":[100,161],"xmlns:xlink=\"http://www.w3.org/1999/xlink\">SemInv</i>":[101,162],"that":[102,171],"consists":[103],"of":[104,132,199],"two":[105],"key":[106],"contributions:":[107],"consistent":[108],"semantics":[109,116],"identifiable":[112,146],"condition":[113,137],"inspection.":[114],"Consistent":[115],"introduces":[118],"new":[120],"regularization":[121],"technique":[122],"into":[123],"the":[124,140,164,197],"optimization":[126],"process,":[127],"enabling":[128],"more":[129,205],"constrained":[134],"Identifiable":[136],"inspection":[138],"assesses":[139],"attack":[141],"performance":[142],"margin":[143],"across":[144],"different":[145],"conditions,":[147],"providing":[148],"robust":[149],"evidence":[150],"distinguishing":[152],"We":[158],"evaluate":[159],"using":[163],"TrojAI":[165],"round":[166],"6\u20138":[167],"datasets":[168],"demonstrate":[170],"significantly":[173],"outperforms":[174],"state-of-the-art":[175],"both":[178],"accuracy":[181],"performance.":[185],"Our":[186],"method":[187],"also":[188],"proves":[189],"against":[191],"with":[193],"stealthy":[194],"triggers,":[195],"advancing":[196],"field":[198],"NLP":[200],"by":[202],"offering":[203],"comprehensive":[206],"solution":[207],"identifying":[209],"attacks.":[211],"The":[212],"code":[213],"repository":[214],"is":[215],"https://github.com/Bluedask/SemInv.":[217]},"counts_by_year":[{"year":2026,"cited_by_count":1}],"updated_date":"2026-03-27T05:58:40.876381","created_date":"2025-10-10T00:00:00"}