{"id":"https://openalex.org/W4415443929","doi":"https://doi.org/10.1109/tdsc.2025.3624268","title":"Alert-ME: An Explainability-Driven Defense Against Adversarial Examples in Transformer-Based Text Classification","display_name":"Alert-ME: An Explainability-Driven Defense Against Adversarial Examples in Transformer-Based Text Classification","publication_year":2025,"publication_date":"2025-10-22","ids":{"openalex":"https://openalex.org/W4415443929","doi":"https://doi.org/10.1109/tdsc.2025.3624268"},"language":null,"primary_location":{"id":"doi:10.1109/tdsc.2025.3624268","is_oa":false,"landing_page_url":"https://doi.org/10.1109/tdsc.2025.3624268","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,"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"},"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/A5089776133","display_name":"Bushra Sabir","orcid":"https://orcid.org/0000-0003-4303-5169"},"institutions":[{"id":"https://openalex.org/I42894916","display_name":"Data61","ror":"https://ror.org/03q397159","country_code":"AU","type":"other","lineage":["https://openalex.org/I1292875679","https://openalex.org/I2801453606","https://openalex.org/I42894916","https://openalex.org/I4387156119"]},{"id":"https://openalex.org/I5681781","display_name":"The University of Adelaide","ror":"https://ror.org/00892tw58","country_code":"AU","type":"education","lineage":["https://openalex.org/I5681781"]}],"countries":["AU"],"is_corresponding":false,"raw_author_name":"Bushra Sabir","raw_affiliation_strings":["CSIROs Data61, Eveleigh, NSW, Australia","School of Computer Science, The University of Adelaide, Australia"],"raw_orcid":"https://orcid.org/0000-0003-4303-5169","affiliations":[{"raw_affiliation_string":"CSIROs Data61, Eveleigh, NSW, Australia","institution_ids":["https://openalex.org/I42894916"]},{"raw_affiliation_string":"School of Computer Science, The University of Adelaide, Australia","institution_ids":["https://openalex.org/I5681781"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101863680","display_name":"Yansong Gao","orcid":"https://orcid.org/0000-0001-5783-2172"},"institutions":[{"id":"https://openalex.org/I177877127","display_name":"The University of Western Australia","ror":"https://ror.org/047272k79","country_code":"AU","type":"education","lineage":["https://openalex.org/I177877127"]},{"id":"https://openalex.org/I4210102982","display_name":"Centre for Research in Engineering Surface Technology","ror":"https://ror.org/018q3fa22","country_code":"IE","type":"other","lineage":["https://openalex.org/I4210102982"]},{"id":"https://openalex.org/I4210159203","display_name":"Western Australian Marine Science Institution","ror":"https://ror.org/05vtzry32","country_code":"AU","type":"facility","lineage":["https://openalex.org/I4210159203"]}],"countries":["AU","IE"],"is_corresponding":false,"raw_author_name":"Yansong Gao","raw_affiliation_strings":["Department of Computer Science, Software Engineering, University of Western Australia, Crawley, WA, Australia","CREST - The Centre for Research on Engineering Software Technologies, Australia"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Department of Computer Science, Software Engineering, University of Western Australia, Crawley, WA, Australia","institution_ids":["https://openalex.org/I177877127","https://openalex.org/I4210159203"]},{"raw_affiliation_string":"CREST - The Centre for Research on Engineering Software Technologies, Australia","institution_ids":["https://openalex.org/I4210102982"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5086357020","display_name":"Alsharif Abuadbba","orcid":"https://orcid.org/0000-0001-9695-7947"},"institutions":[{"id":"https://openalex.org/I1292875679","display_name":"Commonwealth Scientific and Industrial Research Organisation","ror":"https://ror.org/03qn8fb07","country_code":"AU","type":"government","lineage":["https://openalex.org/I1292875679","https://openalex.org/I2801453606","https://openalex.org/I4387156119"]},{"id":"https://openalex.org/I42894916","display_name":"Data61","ror":"https://ror.org/03q397159","country_code":"AU","type":"other","lineage":["https://openalex.org/I1292875679","https://openalex.org/I2801453606","https://openalex.org/I42894916","https://openalex.org/I4387156119"]}],"countries":["AU"],"is_corresponding":false,"raw_author_name":"Alsharif Abuadbba","raw_affiliation_strings":["CSIROs Data61, Eveleigh, NSW, Australia","CSIROs Data61, Australia"],"raw_orcid":"https://orcid.org/0000-0001-9695-7947","affiliations":[{"raw_affiliation_string":"CSIROs Data61, Eveleigh, NSW, Australia","institution_ids":["https://openalex.org/I42894916"]},{"raw_affiliation_string":"CSIROs Data61, Australia","institution_ids":["https://openalex.org/I42894916","https://openalex.org/I1292875679"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5103075476","display_name":"Muhammad Ali Babar","orcid":"https://orcid.org/0000-0001-9696-3626"},"institutions":[{"id":"https://openalex.org/I177877127","display_name":"The University of Western Australia","ror":"https://ror.org/047272k79","country_code":"AU","type":"education","lineage":["https://openalex.org/I177877127"]},{"id":"https://openalex.org/I5681781","display_name":"The University of Adelaide","ror":"https://ror.org/00892tw58","country_code":"AU","type":"education","lineage":["https://openalex.org/I5681781"]}],"countries":["AU"],"is_corresponding":false,"raw_author_name":"M. Ali Babar","raw_affiliation_strings":["School of Computer Science, University of Adelaide, Adelaide, SA, Australia","Deparment of Computer Science and Software Engineering, The University of Western Australia, Australia"],"raw_orcid":"https://orcid.org/0000-0001-9696-3626","affiliations":[{"raw_affiliation_string":"School of Computer Science, University of Adelaide, Adelaide, SA, Australia","institution_ids":["https://openalex.org/I5681781"]},{"raw_affiliation_string":"Deparment of Computer Science and Software Engineering, The University of Western Australia, Australia","institution_ids":["https://openalex.org/I177877127"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":4,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.14229981,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":"23","issue":"2","first_page":"2085","last_page":"2099"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.9894000291824341,"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":0.9894000291824341,"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/adversarial-system","display_name":"Adversarial system","score":0.8004999756813049},{"id":"https://openalex.org/keywords/robustness","display_name":"Robustness (evolution)","score":0.6478000283241272},{"id":"https://openalex.org/keywords/process","display_name":"Process (computing)","score":0.4893999993801117},{"id":"https://openalex.org/keywords/natural-language","display_name":"Natural language","score":0.4742000102996826},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.4708000123500824},{"id":"https://openalex.org/keywords/identification","display_name":"Identification (biology)","score":0.46560001373291016},{"id":"https://openalex.org/keywords/transparency","display_name":"Transparency (behavior)","score":0.44769999384880066},{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.382099986076355}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.848800003528595},{"id":"https://openalex.org/C37736160","wikidata":"https://www.wikidata.org/wiki/Q1801315","display_name":"Adversarial system","level":2,"score":0.8004999756813049},{"id":"https://openalex.org/C63479239","wikidata":"https://www.wikidata.org/wiki/Q7353546","display_name":"Robustness (evolution)","level":3,"score":0.6478000283241272},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5472999811172485},{"id":"https://openalex.org/C98045186","wikidata":"https://www.wikidata.org/wiki/Q205663","display_name":"Process (computing)","level":2,"score":0.4893999993801117},{"id":"https://openalex.org/C195324797","wikidata":"https://www.wikidata.org/wiki/Q33742","display_name":"Natural language","level":2,"score":0.4742000102996826},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.4708000123500824},{"id":"https://openalex.org/C116834253","wikidata":"https://www.wikidata.org/wiki/Q2039217","display_name":"Identification (biology)","level":2,"score":0.46560001373291016},{"id":"https://openalex.org/C2780233690","wikidata":"https://www.wikidata.org/wiki/Q535347","display_name":"Transparency (behavior)","level":2,"score":0.44769999384880066},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.39320001006126404},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.382099986076355},{"id":"https://openalex.org/C43521106","wikidata":"https://www.wikidata.org/wiki/Q2165493","display_name":"Pipeline (software)","level":2,"score":0.3531000018119812},{"id":"https://openalex.org/C95713431","wikidata":"https://www.wikidata.org/wiki/Q631425","display_name":"Vulnerability (computing)","level":2,"score":0.33730000257492065},{"id":"https://openalex.org/C38652104","wikidata":"https://www.wikidata.org/wiki/Q3510521","display_name":"Computer security","level":1,"score":0.301800012588501},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.29589998722076416},{"id":"https://openalex.org/C124304363","wikidata":"https://www.wikidata.org/wiki/Q673661","display_name":"Abstraction","level":2,"score":0.28940001130104065},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.2847999930381775},{"id":"https://openalex.org/C44359876","wikidata":"https://www.wikidata.org/wiki/Q5338467","display_name":"Edit distance","level":2,"score":0.2689000070095062},{"id":"https://openalex.org/C108583219","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep learning","level":2,"score":0.2630000114440918},{"id":"https://openalex.org/C2780801425","wikidata":"https://www.wikidata.org/wiki/Q5164392","display_name":"Construct (python library)","level":2,"score":0.2590000033378601},{"id":"https://openalex.org/C2779791154","wikidata":"https://www.wikidata.org/wiki/Q258040","display_name":"Model transformation","level":3,"score":0.2563999891281128},{"id":"https://openalex.org/C90805587","wikidata":"https://www.wikidata.org/wiki/Q10944557","display_name":"Word (group theory)","level":2,"score":0.25380000472068787},{"id":"https://openalex.org/C52622490","wikidata":"https://www.wikidata.org/wiki/Q1026626","display_name":"Feature extraction","level":2,"score":0.25279998779296875}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/tdsc.2025.3624268","is_oa":false,"landing_page_url":"https://doi.org/10.1109/tdsc.2025.3624268","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,"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":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":44,"referenced_works":["https://openalex.org/W2081580037","https://openalex.org/W2166391512","https://openalex.org/W2250539671","https://openalex.org/W2282821441","https://openalex.org/W2493916176","https://openalex.org/W2800788706","https://openalex.org/W2883364792","https://openalex.org/W2905526464","https://openalex.org/W2918408501","https://openalex.org/W2949128310","https://openalex.org/W2962818281","https://openalex.org/W2963341956","https://openalex.org/W2973136764","https://openalex.org/W2979826702","https://openalex.org/W2982054702","https://openalex.org/W2996851481","https://openalex.org/W2999309192","https://openalex.org/W3035441470","https://openalex.org/W3085753079","https://openalex.org/W3101449015","https://openalex.org/W3104423855","https://openalex.org/W3105604018","https://openalex.org/W3127569499","https://openalex.org/W3155936402","https://openalex.org/W3168586460","https://openalex.org/W3169965252","https://openalex.org/W3176196997","https://openalex.org/W3197868468","https://openalex.org/W3204619801","https://openalex.org/W3213226566","https://openalex.org/W4211137019","https://openalex.org/W4230405732","https://openalex.org/W4239025696","https://openalex.org/W4251372957","https://openalex.org/W4285049279","https://openalex.org/W4285116762","https://openalex.org/W4285204696","https://openalex.org/W4285266418","https://openalex.org/W4285550464","https://openalex.org/W4313196990","https://openalex.org/W4321786089","https://openalex.org/W4383221565","https://openalex.org/W4385245566","https://openalex.org/W4392849751"],"related_works":[],"abstract_inverted_index":{"Transformer-based":[0],"Text":[1],"Classifiers":[2],"(TTC),":[3],"exemplified":[4],"by":[5,299],"models":[6],"like":[7,267],"BERT,":[8],"Roberta,":[9],"T5,":[10],"and":[11,48,66,89,95,126,148,159,171,183,200,213,218,224,242,270,277,287,297,302,351],"GPT,":[12],"have":[13],"showcased":[14],"impressive":[15],"proficiency":[16],"in":[17,238,246,251,259],"Natural":[18],"Language":[19],"Processing":[20],"(NLP).":[21],"Nonetheless,":[22],"their":[23],"vulnerability":[24],"to":[25,36,71,129,155,227],"adversarial":[26,115,311],"examples":[27,116,312],"(AEs)":[28],"presents":[29],"a":[30,53,152],"substantial":[31],"security":[32,153],"concern.":[33],"Existing":[34],"efforts":[35],"fortify":[37],"model":[38,127],"robustness":[39],"either":[40],"entail":[41],"high":[42],"computational":[43,358],"costs":[44],"or":[45],"lack":[46],"transparency":[47],"interpretability.":[49],"This":[50],"paper":[51],"introduces":[52],"holistic":[54],"framework":[55,144,168,189],"called":[56],"<underline":[57,60,63,67],"xmlns:mml=\"http://www.w3.org/1998/Math/MathML\"":[58,61,64,68],"xmlns:xlink=\"http://www.w3.org/1999/xlink\">E</u>xplainability-driven":[59],"xmlns:xlink=\"http://www.w3.org/1999/xlink\">D</u>etection,":[62],"xmlns:xlink=\"http://www.w3.org/1999/xlink\">I</u>dentification,":[65],"xmlns:xlink=\"http://www.w3.org/1999/xlink\">T</u>ransformation":[69],"(EDIT)":[70],"address":[72],"the":[73,104,131,143,146,323],"shortcomings":[74],"of":[75,97,202,216,319,333,344,353],"prevailing":[76],"inference-time":[77],"defenses.":[78],"(i)":[79],"EDIT":[80,112,208,321,338],"employs":[81],"explainability":[82],"methodologies":[83],"(e.g.,":[84],"attention":[85],"maps,":[86],"integrated":[87],"gradients)":[88],"frequency":[90],"features":[91],"for":[92,134,151,325],"automated":[93],"detection":[94],"identification":[96,263],"perturbed":[98,136],"words":[99],"while":[100,356],"providing":[101],"insights":[102],"into":[103,313],"model's":[105],"decision-making":[106],"process.":[107],"(ii)":[108],"Building":[109],"upon":[110],"detection,":[111],"purifies":[113],"detected":[114],"through":[117],"an":[118,210,317,329,340],"optimal":[119],"transformation":[120,307],"mechanism":[121],"that":[122,207],"utilizes":[123],"pre-trained":[124],"embeddings":[125],"feedback":[128],"find":[130],"best":[132],"replacements":[133],"identified":[135],"words.":[137],"(iii)":[138],"To":[139],"integrate":[140],"human":[141,162,326],"intelligence,":[142],"automates":[145],"process":[147],"triggers":[149],"alerts":[150,336],"analyst":[154],"ensure":[156],"secure":[157],"safeguards":[158],"enable":[160],"proactive":[161],"feedback.":[163],"We":[164,185],"comprehensively":[165],"evaluate":[166],"our":[167,188,232],"on":[169,174],"BERT":[170],"ROBERTA,":[172],"trained":[173],"four":[175,228],"SOTA":[176,192,229],"text":[177],"classification":[178],"datasets:":[179],"IMDB,":[180],"YELP,":[181],"AGNEWS,":[182],"SST2.":[184],"also":[186,249],"assess":[187],"against":[190],"seven":[191],"Word":[193],"Substitution":[194],"Attacks":[195],"(WSA)":[196],"at":[197],"character,":[198],"word,":[199],"multi-levels":[201],"granularity.":[203],"The":[204,262],"results":[205],"show":[206],"achieves":[209,339],"average":[211,330,341],"F-score":[212],"balanced":[214,239],"accuracy":[215,240,318,332,343],"89.69%":[217],"89.70%,":[219],"respectively,":[220],"across":[221,346],"all":[222,347,355],"datasets":[223],"classifiers.":[225],"Compared":[226],"inference":[230],"defenses,":[231],"detector":[233],"is":[234],"1.22":[235],"times":[236,244,257,274,280,284,289],"better":[237,245,275],"(BAL_ACC)":[241],"1.33":[243],"F1-Score.":[247,291],"It":[248,292],"excels":[250],"operational":[252],"efficiency,":[253,359],"being":[254],"about":[255],"83":[256],"faster":[258],"feature":[260],"extraction.":[261],"module":[264,308],"outperforms":[265],"techniques":[266],"ReplaceScore,":[268],"FreqScore,":[269],"ExplainScore,":[271],"achieving":[272],"1.125":[273],"BAL_ACC":[276],"Recall,":[278],"1.142":[279],"higher":[281],"AUC,":[282],"1.120":[283],"greater":[285],"Precision,":[286],"1.149":[288],"improved":[290],"significantly":[293],"reduces":[294],"false":[295],"positives":[296],"negatives":[298],"approximately":[300],"1.537":[301],"1.172":[303],"times,":[304],"respectively.":[305],"Our":[306],"effectively":[309],"converts":[310],"non-adversarial":[314],"counterparts":[315],"with":[316,328],"91%.":[320],"identifies":[322],"necessity":[324],"intervention":[327],"median":[331,342],"89%.":[334],"With":[335],"integrated,":[337],"90%":[345],"evaluated":[348],"datasets,":[349],"models,":[350],"types":[352],"attacks,":[354],"maintaining":[357],"averaging":[360],"6.94":[361],"seconds.":[362]},"counts_by_year":[],"updated_date":"2026-06-11T09:08:48.828518","created_date":"2025-10-24T00:00:00"}