{"id":"https://openalex.org/W7105650856","doi":"https://doi.org/10.1109/ijcnn64981.2025.11229281","title":"DEO: Jailbreak a Black-box Multimodal Large Language Model with Dual-Embedding Alignment","display_name":"DEO: Jailbreak a Black-box Multimodal Large Language Model with Dual-Embedding Alignment","publication_year":2025,"publication_date":"2025-06-30","ids":{"openalex":"https://openalex.org/W7105650856","doi":"https://doi.org/10.1109/ijcnn64981.2025.11229281"},"language":null,"primary_location":{"id":"doi:10.1109/ijcnn64981.2025.11229281","is_oa":false,"landing_page_url":"https://doi.org/10.1109/ijcnn64981.2025.11229281","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 International Joint Conference on Neural Networks (IJCNN)","raw_type":"proceedings-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":null,"display_name":"Lijie Zhang","orcid":null},"institutions":[{"id":"https://openalex.org/I19820366","display_name":"Chinese Academy of Sciences","ror":"https://ror.org/034t30j35","country_code":"CN","type":"funder","lineage":["https://openalex.org/I19820366"]},{"id":"https://openalex.org/I4210156404","display_name":"Institute of Information Engineering","ror":"https://ror.org/04r53se39","country_code":"CN","type":"facility","lineage":["https://openalex.org/I19820366","https://openalex.org/I4210156404"]}],"countries":["CN"],"is_corresponding":true,"raw_author_name":"Lijie Zhang","raw_affiliation_strings":["Chinese Academy of Sciences,Institute of Information Engineering,China"],"affiliations":[{"raw_affiliation_string":"Chinese Academy of Sciences,Institute of Information Engineering,China","institution_ids":["https://openalex.org/I4210156404","https://openalex.org/I19820366"]}]},{"author_position":"middle","author":{"id":null,"display_name":"Mingsi Wang","orcid":null},"institutions":[{"id":"https://openalex.org/I4210156404","display_name":"Institute of Information Engineering","ror":"https://ror.org/04r53se39","country_code":"CN","type":"facility","lineage":["https://openalex.org/I19820366","https://openalex.org/I4210156404"]},{"id":"https://openalex.org/I19820366","display_name":"Chinese Academy of Sciences","ror":"https://ror.org/034t30j35","country_code":"CN","type":"funder","lineage":["https://openalex.org/I19820366"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Mingsi Wang","raw_affiliation_strings":["Chinese Academy of Sciences,Institute of Information Engineering,China"],"affiliations":[{"raw_affiliation_string":"Chinese Academy of Sciences,Institute of Information Engineering,China","institution_ids":["https://openalex.org/I4210156404","https://openalex.org/I19820366"]}]},{"author_position":"middle","author":{"id":null,"display_name":"Yue Zhao","orcid":null},"institutions":[{"id":"https://openalex.org/I19820366","display_name":"Chinese Academy of Sciences","ror":"https://ror.org/034t30j35","country_code":"CN","type":"funder","lineage":["https://openalex.org/I19820366"]},{"id":"https://openalex.org/I4210156404","display_name":"Institute of Information Engineering","ror":"https://ror.org/04r53se39","country_code":"CN","type":"facility","lineage":["https://openalex.org/I19820366","https://openalex.org/I4210156404"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Yue Zhao","raw_affiliation_strings":["Chinese Academy of Sciences,Institute of Information Engineering,China"],"affiliations":[{"raw_affiliation_string":"Chinese Academy of Sciences,Institute of Information Engineering,China","institution_ids":["https://openalex.org/I4210156404","https://openalex.org/I19820366"]}]},{"author_position":"middle","author":{"id":null,"display_name":"Zijin Lin","orcid":null},"institutions":[{"id":"https://openalex.org/I4210156404","display_name":"Institute of Information Engineering","ror":"https://ror.org/04r53se39","country_code":"CN","type":"facility","lineage":["https://openalex.org/I19820366","https://openalex.org/I4210156404"]},{"id":"https://openalex.org/I19820366","display_name":"Chinese Academy of Sciences","ror":"https://ror.org/034t30j35","country_code":"CN","type":"funder","lineage":["https://openalex.org/I19820366"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Zijin Lin","raw_affiliation_strings":["Chinese Academy of Sciences,Institute of Information Engineering,China"],"affiliations":[{"raw_affiliation_string":"Chinese Academy of Sciences,Institute of Information Engineering,China","institution_ids":["https://openalex.org/I4210156404","https://openalex.org/I19820366"]}]},{"author_position":"last","author":{"id":null,"display_name":"Kai Chen","orcid":null},"institutions":[{"id":"https://openalex.org/I19820366","display_name":"Chinese Academy of Sciences","ror":"https://ror.org/034t30j35","country_code":"CN","type":"funder","lineage":["https://openalex.org/I19820366"]},{"id":"https://openalex.org/I4210156404","display_name":"Institute of Information Engineering","ror":"https://ror.org/04r53se39","country_code":"CN","type":"facility","lineage":["https://openalex.org/I19820366","https://openalex.org/I4210156404"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Kai Chen","raw_affiliation_strings":["Chinese Academy of Sciences,Institute of Information Engineering,China"],"affiliations":[{"raw_affiliation_string":"Chinese Academy of Sciences,Institute of Information Engineering,China","institution_ids":["https://openalex.org/I4210156404","https://openalex.org/I19820366"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":5,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I19820366","https://openalex.org/I4210156404"],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.79344181,"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":"8"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.6740999817848206,"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":0.6740999817848206,"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/T10775","display_name":"Generative Adversarial Networks and Image Synthesis","score":0.07050000131130219,"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/T11714","display_name":"Multimodal Machine Learning Applications","score":0.03869999945163727,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/embedding","display_name":"Embedding","score":0.842199981212616},{"id":"https://openalex.org/keywords/adversarial-system","display_name":"Adversarial system","score":0.7914999723434448},{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.5278000235557556},{"id":"https://openalex.org/keywords/key","display_name":"Key (lock)","score":0.3772999942302704},{"id":"https://openalex.org/keywords/exploit","display_name":"Exploit","score":0.33009999990463257},{"id":"https://openalex.org/keywords/visualization","display_name":"Visualization","score":0.31279999017715454}],"concepts":[{"id":"https://openalex.org/C41608201","wikidata":"https://www.wikidata.org/wiki/Q980509","display_name":"Embedding","level":2,"score":0.842199981212616},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8122000098228455},{"id":"https://openalex.org/C37736160","wikidata":"https://www.wikidata.org/wiki/Q1801315","display_name":"Adversarial system","level":2,"score":0.7914999723434448},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.5278000235557556},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.48899999260902405},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.3772999942302704},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.3343000113964081},{"id":"https://openalex.org/C165696696","wikidata":"https://www.wikidata.org/wiki/Q11287","display_name":"Exploit","level":2,"score":0.33009999990463257},{"id":"https://openalex.org/C36464697","wikidata":"https://www.wikidata.org/wiki/Q451553","display_name":"Visualization","level":2,"score":0.31279999017715454},{"id":"https://openalex.org/C2780878386","wikidata":"https://www.wikidata.org/wiki/Q1659648","display_name":"Visual language","level":2,"score":0.3077999949455261},{"id":"https://openalex.org/C107457646","wikidata":"https://www.wikidata.org/wiki/Q207434","display_name":"Human\u2013computer interaction","level":1,"score":0.2939999997615814},{"id":"https://openalex.org/C140547941","wikidata":"https://www.wikidata.org/wiki/Q7797194","display_name":"Threat model","level":2,"score":0.290800005197525},{"id":"https://openalex.org/C2983448237","wikidata":"https://www.wikidata.org/wiki/Q1078276","display_name":"Language understanding","level":2,"score":0.2881999909877777},{"id":"https://openalex.org/C31972630","wikidata":"https://www.wikidata.org/wiki/Q844240","display_name":"Computer vision","level":1,"score":0.2678000032901764},{"id":"https://openalex.org/C115961682","wikidata":"https://www.wikidata.org/wiki/Q860623","display_name":"Image (mathematics)","level":2,"score":0.2669000029563904},{"id":"https://openalex.org/C195324797","wikidata":"https://www.wikidata.org/wiki/Q33742","display_name":"Natural language","level":2,"score":0.26019999384880066}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/ijcnn64981.2025.11229281","is_oa":false,"landing_page_url":"https://doi.org/10.1109/ijcnn64981.2025.11229281","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 International Joint Conference on Neural Networks (IJCNN)","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[{"display_name":"Peace, Justice and strong institutions","score":0.5873537659645081,"id":"https://metadata.un.org/sdg/16"}],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":15,"referenced_works":["https://openalex.org/W2149479912","https://openalex.org/W3098267758","https://openalex.org/W3158360872","https://openalex.org/W4312605942","https://openalex.org/W4312933868","https://openalex.org/W4392353733","https://openalex.org/W4393157467","https://openalex.org/W4402727764","https://openalex.org/W4403878024","https://openalex.org/W4404356490","https://openalex.org/W4404971291","https://openalex.org/W4409348010","https://openalex.org/W4410609100","https://openalex.org/W4411799673","https://openalex.org/W4413277723"],"related_works":[],"abstract_inverted_index":{"Multimodal":[0],"Large":[1],"Language":[2],"Models":[3],"(MLLMs),":[4],"which":[5,32,157],"integrate":[6],"textual":[7],"and":[8,124,134,183,214,249,262,280],"visual":[9,23,47,92,114],"modalities,":[10],"have":[11,39,53],"demonstrated":[12],"unparalleled":[13],"capabilities":[14],"in":[15,50,66,74,199,284],"diverse":[16],"multimodal":[17],"tasks.":[18],"However,":[19],"the":[20,46,97,113,122,128,132,135,140,179,186,200,218,273],"inclusion":[21],"of":[22,31,131,220,234,257],"inputs":[24],"exposes":[25],"MLLMs":[26,44,98],"to":[27,42,61,90,99,145,191,210,241],"security":[28],"risks,":[29],"one":[30],"is":[33,158,167],"jailbreak":[34,43],"attacks.":[35],"Although":[36],"various":[37],"methods":[38,238],"been":[40],"proposed":[41],"via":[45],"modality,":[48],"attacks":[49,58],"black-box":[51,57,172,236],"settings":[52],"some":[54],"limitations.":[55],"Existing":[56],"either":[59],"fail":[60],"generate":[62,91,206],"precise":[63,215],"harmful":[64,101,149,213,292],"outputs":[65],"practical":[67],"scenarios":[68],"or":[69,195],"require":[70],"substantial":[71],"preparatory":[72],"work":[73],"constructing":[75],"adversarial":[76,93,208],"images.":[77],"In":[78],"this":[79],"work,":[80],"we":[81],"propose":[82],"a":[83,148,153,161,171,175],"novel":[84],"dual-embedding":[85,201],"optimization":[86],"(DEO)":[87],"attack":[88,231,237,254,268],"approach":[89],"perturbations":[94],"that":[95,103,226],"induce":[96],"produce":[100],"responses":[102],"violate":[104],"common":[105],"AI":[106],"safety":[107,282],"policies.":[108],"Specifically,":[109],"DEO":[110],"iteratively":[111],"optimizes":[112],"input":[115,123,133],"by":[116,139,160,239],"enforcing":[117],"alignment":[118,166],"objectives":[119],"across":[120,259],"both":[121,143],"output":[125],"embedding":[126,130,137,155],"spaces:":[127],"image":[129],"text":[136,151],"generated":[138],"MLLM":[141,245],"are":[142],"required":[144],"align":[146],"with":[147],"target":[150],"within":[152],"shared":[154],"space,":[156,202],"defined":[159],"frozen":[162],"pretrained":[163],"encoder.":[164],"This":[165,289],"conducted":[168],"entirely":[169],"under":[170],"setting":[173],"using":[174],"query-based":[176],"strategy,":[177],"where":[178],"attacker":[180],"issues":[181],"queries":[182],"observes":[184],"only":[185],"model\u2019s":[187],"outputs,":[188],"without":[189],"access":[190],"its":[192,266],"internal":[193],"parameters":[194],"gradients.":[196],"By":[197],"optimizing":[198],"our":[203,227],"method":[204,228],"can":[205],"an":[207,252],"perturbation":[209],"elicit":[211],"more":[212],"responses,":[216],"overcoming":[217],"limitations":[219],"existing":[221,235],"approaches.":[222],"Experimental":[223],"results":[224],"demonstrate":[225],"significantly":[229],"improves":[230],"success":[232,255],"rates":[233],"up":[240],"30%":[242],"against":[243],"two":[244],"families,":[246],"including":[247],"MiniGPT4":[248],"LLaVa,":[250],"achieving":[251],"average":[253],"rate":[256],"87%":[258],"different":[260],"models":[261],"eight":[263],"scenarios,":[264],"demonstrating":[265],"superior":[267],"effectiveness.":[269],"These":[270],"findings":[271],"highlight":[272],"urgent":[274],"need":[275],"for":[276],"systematic":[277],"robustness":[278],"evaluations":[279],"improved":[281],"mechanisms":[283],"MLLMs.<sup":[285],"xmlns:mml=\"http://www.w3.org/1998/Math/MathML\"":[286],"xmlns:xlink=\"http://www.w3.org/1999/xlink\">1</sup>Content":[287],"Warning:":[288],"paper":[290],"contains":[291],"model":[293],"responses.":[294]},"counts_by_year":[],"updated_date":"2025-11-15T23:13:30.683059","created_date":"2025-11-14T00:00:00"}