{"authors":[{"id":null,"fullName":"Zarras, Apostolis","name":"Apostolis","surname":"Zarras","rank":1,"pid":{"id":{"scheme":"orcid_pending","value":"0000-0002-8480-6989"},"provenance":null}},{"id":"orcid_______::9f924a205fb7188f2ae9d04a414d8c48","fullName":"Kollarou, Athanasia","name":"Athanasia","surname":"Kollarou","rank":2,"pid":{"id":{"scheme":"orcid","value":"0009-0008-8239-4432"},"provenance":null}},{"id":null,"fullName":"Farao, Aristeidis","name":"Aristeidis","surname":"Farao","rank":3,"pid":{"id":{"scheme":"orcid_pending","value":"0000-0001-6954-0791"},"provenance":null}},{"id":null,"fullName":"Bountakas, Panagiotis","name":"Panagiotis","surname":"Bountakas","rank":4,"pid":{"id":{"scheme":"orcid_pending","value":"0000-0001-8155-1784"},"provenance":null}},{"id":null,"fullName":"Xenakis, Christos","name":"Christos","surname":"Xenakis","rank":5,"pid":{"id":{"scheme":"orcid_pending","value":"0000-0001-6718-122x"},"provenance":null}}],"openAccessColor":"gold","publiclyFunded":false,"eoscIfGuidelines":null,"type":"publication","language":{"code":"eng","label":"English"},"countries":null,"subjects":[{"subject":{"scheme":"FOS","value":"0202 electrical engineering, electronic engineering, information engineering"},"provenance":null},{"subject":{"scheme":"FOS","value":"02 engineering and technology"},"provenance":null}],"mainTitle":"Testing the limits: exploring adversarial techniques in AI models","subTitle":null,"descriptions":["<jats:p>                     The rising adoption of artificial intelligence and machine learning in critical sectors underscores the pressing need for robust systems capable of withstanding adversarial threats. While deep learning architectures have revolutionized tasks such as image recognition, their susceptibility to adversarial techniques remains an open challenge. This article evaluates the impact of various adversarial methods, including the fast gradient sign method, projected gradient descent, DeepFool, and Carlini &amp; Wagner, on five neural network models: a fully connected neural network, LeNet, Simple convolutional neural network (CNN), MobileNetV2, and VGG11. Using the E                     <jats:sc>V</jats:sc>                     AI                     <jats:sc>SION</jats:sc>                     tool explicitly developed for this research, these attacks were implemented and analyzed based on accuracy, F1-score, and misclassification rate. The results revealed varying levels of vulnerability across the tested models, with simpler architectures occasionally outperforming more complex ones. These findings emphasize the importance of selecting the most appropriate adversarial technique for a given architecture and customizing the associated attack parameters to achieve optimal results in each scenario.                   </jats:p>"],"publicationDate":"2025-10-31","publisher":"PeerJ","embargoEndDate":null,"sources":["Crossref","PeerJ Computer Science"],"formats":null,"contributors":null,"coverages":null,"bestAccessRight":{"code":"c_abf2","label":"OPEN","scheme":"http://vocabularies.coar-repositories.org/documentation/access_rights/"},"container":{"name":"PeerJ Computer Science","issnPrinted":null,"issnOnline":"2376-5992","issnLinking":null,"ep":null,"iss":null,"sp":"e3330","vol":"11","edition":null,"conferencePlace":null,"conferenceDate":null},"documentationUrls":null,"codeRepositoryUrl":null,"programmingLanguage":null,"contactPeople":null,"contactGroups":null,"tools":null,"size":null,"version":null,"geoLocations":null,"id":"doi_dedup___::e08eccc8d2ff06da3b0d085e956c4c11","originalIds":["10.7717/peerj-cs.3330","50|doiboost____|e08eccc8d2ff06da3b0d085e956c4c11","50|dblp________::3ece628f93e6fc745f3c116bd625011a","oai:zenodo.org:17708030","50|od______2659::e08eccc8d2ff06da3b0d085e956c4c11","10.7717/PEERJ-CS.3330","50|r3c4b2081b22::f683283d55869e9976fec8e61710e04c","50|sygma_______::e08eccc8d2ff06da3b0d085e956c4c11"],"pids":[{"scheme":"doi","value":"10.7717/peerj-cs.3330"}],"dateOfCollection":null,"lastUpdateTimeStamp":null,"indicators":{"citationImpact":{"citationCount":0.0,"influence":2.2251732E-9,"popularity":2.342466E-9,"impulse":0.0,"citationClass":"C5","influenceClass":"C5","impulseClass":"C5","popularityClass":"C5"}},"projects":[{"id":"corda_____he::3c2f079846d658cf109b663d00c966e2","code":"101183162","acronym":"ANTIDOTE","title":"AI Attack and Defense for the Smart Healthcare","funder":"European Commission","pids":[{"scheme":"doi","value":"10.3030/101183162"}]},{"id":"corda_____he::a6424752830a86a19f2d111c98c462f6","code":"101120962","acronym":"RESCALE","title":"Revolutionised Enhanced Supply Chain Automation with Limited Threats Exposure","funder":"European Commission","pids":[{"scheme":"doi","value":"10.3030/101120962"}]},{"id":"corda_____he::805f2a5d936f1e1aaaa3e1f216b026f8","code":"101131292","acronym":"AIAS","title":"AI-ASsisted cybersecurity platform empowering SMEs to defend against adversarial AI attacks","funder":"European Commission","pids":[{"scheme":"doi","value":"10.3030/101131292"}]},{"id":"corda_____he::7868501db8c07a797fd0de7817d6a54b","code":"101168407","acronym":"cPAID","title":"Cloud-based Platform-agnostic Adversarial aI Defence framework– CPAID","funder":"European Commission","pids":[{"scheme":"doi","value":"10.3030/101168407"}]}],"organizations":[{"legalName":"Foundation for Research and Technology Hellas","acronym":"FOUNDATION FOR RESEARCH AND TECHNOLOGYHELLAS","id":"openorgs____::5ec3d36c57c96bdfd6ee8183fb247a90","pids":[{"scheme":"fundref","value":"501100012288"},{"scheme":"GRID","value":"grid.4834.b"},{"scheme":"Wikidata","value":"Q56649061"},{"scheme":"FundRef","value":"501100012288"},{"scheme":"OrgRef","value":"2696935"},{"scheme":"PIC","value":"999995893"},{"scheme":"ROR","value":"https://ror.org/052rphn09"},{"scheme":"OrgReg","value":"GR1009"},{"scheme":"mag_id","value":"8901234"},{"scheme":"Wikidata","value":"Q39633068"},{"scheme":"ISNI","value":"000000040635685X"},{"scheme":"wikidata","value":"Q953914"},{"scheme":"Wikidata","value":"Q953914"},{"scheme":"wikidata","value":"Q39633068"}]},{"legalName":"University of Piraeus","acronym":"University of Piraeus","id":"openorgs____::84f0c5f5dbb6daf42748485924efde4b","pids":[{"scheme":"OrgRef","value":"18715318"},{"scheme":"OrgReg","value":"GR0011"},{"scheme":"ISNI","value":"0000000105588585"},{"scheme":"Wikidata","value":"Q550263"},{"scheme":"fundref","value":"501100021534"},{"scheme":"mag_id","value":"154757721"},{"scheme":"FundRef","value":"501100021534"},{"scheme":"ROR","value":"https://ror.org/02qs84g94"},{"scheme":"GRID","value":"grid.4463.5"},{"scheme":"wikidata","value":"Q550263"}]},{"legalName":"INQBIT INNOVATIONS SRL","acronym":"INQBIT INNOVATIONS SRL","id":"pending_org_::3ac89377dce88ae8e9bb371fc6938ef1","pids":[{"scheme":"PIC","value":"895370529"}]},{"legalName":"UNIVERSITY OF PIRAEUS RESEARCH CENTER","acronym":"UPRC","id":"openorgs____::ed3cbcb588d8289cee7145d83ed393d2","pids":[{"scheme":"PIC","value":"999586941"}]}],"communities":[{"code":"eosc","label":"EOSC","provenance":null}],"collectedFrom":[{"key":"openaire____::a8db6f6b2ce4fe72e8b2314a9a93e7d9","value":"Sygma"},{"key":"opendoar____::358aee4cc897452c00244351e4d91f69","value":"ZENODO"},{"key":"openaire____::081b82f96300b6a6e3d282bad31cb6e2","value":"Crossref"},{"key":"openaire____::d8b68cc0a53121f6f896883a7d60c1db","value":"DBLP"},{"key":"re3data_____::c4b2081b224be6b3e79d0e5e5556f631","value":"European Union Open Data Portal"}],"instances":[{"pids":[{"scheme":"doi","value":"10.7717/peerj-cs.3330"}],"license":"CC BY","accessRight":{"code":"c_abf2","label":"OPEN","scheme":"http://vocabularies.coar-repositories.org/documentation/access_rights/","openAccessRoute":"gold"},"type":"Article","urls":["https://doi.org/10.7717/peerj-cs.3330"],"publicationDate":"2025-10-31","refereed":"peerReviewed","hostedBy":{"key":"doajarticles::8ebf85347f4d37e55843ff28f4d34f97","value":"PeerJ Computer Science"},"collectedFrom":{"key":"openaire____::081b82f96300b6a6e3d282bad31cb6e2","value":"Crossref"}},{"alternateIdentifiers":[{"scheme":"doi","value":"10.7717/peerj-cs.3330"}],"type":"Article","urls":["https://doi.org/10.7717/peerj-cs.3330"],"publicationDate":"2025-01-01","refereed":"nonPeerReviewed","hostedBy":{"key":"openaire____::d8b68cc0a53121f6f896883a7d60c1db","value":"DBLP"},"collectedFrom":{"key":"openaire____::d8b68cc0a53121f6f896883a7d60c1db","value":"DBLP"}},{"alternateIdentifiers":[{"scheme":"oai","value":"oai:zenodo.org:17708030"},{"scheme":"doi","value":"10.7717/peerj-cs.3330"}],"license":"CC BY","accessRight":{"code":"c_abf2","label":"OPEN","scheme":"http://vocabularies.coar-repositories.org/documentation/access_rights/","openAccessRoute":null},"type":"Article","urls":["https://zenodo.org/records/17708030","http://dx.doi.org/10.7717/peerj-cs.3330"],"publicationDate":"2025-10-31","refereed":"nonPeerReviewed","hostedBy":{"key":"opendoar____::358aee4cc897452c00244351e4d91f69","value":"ZENODO"},"collectedFrom":{"key":"opendoar____::358aee4cc897452c00244351e4d91f69","value":"ZENODO"}},{"alternateIdentifiers":[{"scheme":"doi","value":"10.7717/peerj-cs.3330"}],"accessRight":{"code":"c_abf2","label":"OPEN","scheme":"http://vocabularies.coar-repositories.org/documentation/access_rights/","openAccessRoute":"gold"},"type":"Article","urls":["http://dx.doi.org/10.7717/PEERJ-CS.3330"],"publicationDate":"2025-01-01","refereed":"peerReviewed","hostedBy":{"key":"doajarticles::8ebf85347f4d37e55843ff28f4d34f97","value":"PeerJ Computer Science"},"collectedFrom":{"key":"re3data_____::c4b2081b224be6b3e79d0e5e5556f631","value":"European Union Open Data Portal"}},{"alternateIdentifiers":[{"scheme":"doi","value":"10.7717/peerj-cs.3330"}],"license":"CC BY","accessRight":{"code":"c_abf2","label":"OPEN","scheme":"http://vocabularies.coar-repositories.org/documentation/access_rights/","openAccessRoute":"gold"},"type":"Article","urls":["http://dx.doi.org/10.7717/peerj-cs.3330"],"refereed":"nonPeerReviewed","hostedBy":{"key":"doajarticles::8ebf85347f4d37e55843ff28f4d34f97","value":"PeerJ Computer Science"},"collectedFrom":{"key":"openaire____::a8db6f6b2ce4fe72e8b2314a9a93e7d9","value":"Sygma"}}],"links":[{"header":{"relationType":"resultProject","relationClass":"isProducedBy","relatedIdentifier":"corda_____he::3c2f079846d658cf109b663d00c966e2","relatedRecordType":"project","relationProvenance":"sysimport:crosswalk:repository","trust":"0.9"},"collectedfrom":[{"dsId":"openaire____::3f264f93cf3b0cfc4ede188a6300455c","dsName":"CORDA - COmmon Research DAta Warehouse - Horizon Europe"}],"projectTitle":"AI Attack and Defense for the Smart Healthcare","code":"101183162","funding":{"funder":{"id":"ec__________::EC","shortname":"EC","name":"European Commission","jurisdiction":{"code":"EU","label":"European Union"},"pid":null},"level0":{"id":"ec__________::EC::HE","description":"Horizon Europe Framework Programme","name":"HE"},"level1":{"id":"ec__________::EC::HE::HORIZON-TMA-MSCA-SE","description":"HORIZON TMA MSCA Staff Exchanges","name":"HORIZON-TMA-MSCA-SE"},"level2":{"id":null,"description":null,"name":null}},"startDate":"2025-01-01","endDate":"2028-12-31"},{"header":{"relationType":"resultProject","relationClass":"isProducedBy","relatedIdentifier":"corda_____he::a6424752830a86a19f2d111c98c462f6","relatedRecordType":"project","relationProvenance":"sysimport:crosswalk:repository","trust":"0.9"},"collectedfrom":[{"dsId":"openaire____::3f264f93cf3b0cfc4ede188a6300455c","dsName":"CORDA - COmmon Research DAta Warehouse - Horizon Europe"}],"projectTitle":"Revolutionised Enhanced Supply Chain Automation with Limited Threats Exposure","code":"101120962","funding":{"funder":{"id":"ec__________::EC","shortname":"EC","name":"European Commission","jurisdiction":{"code":"EU","label":"European Union"},"pid":null},"level0":{"id":"ec__________::EC::HE","description":"Horizon Europe Framework Programme","name":"HE"},"level1":{"id":"ec__________::EC::HE::HORIZON-RIA","description":"HORIZON  Research and Innovation Actions","name":"HORIZON-RIA"},"level2":{"id":null,"description":null,"name":null}},"startDate":"2023-10-01","endDate":"2026-09-30"},{"header":{"relationType":"resultProject","relationClass":"isProducedBy","relatedIdentifier":"corda_____he::805f2a5d936f1e1aaaa3e1f216b026f8","relatedRecordType":"project","relationProvenance":"sysimport:crosswalk:repository","trust":"0.9"},"collectedfrom":[{"dsId":"openaire____::3f264f93cf3b0cfc4ede188a6300455c","dsName":"CORDA - COmmon Research DAta Warehouse - Horizon Europe"}],"projectTitle":"AI-ASsisted cybersecurity platform empowering SMEs to defend against adversarial AI attacks","code":"101131292","funding":{"funder":{"id":"ec__________::EC","shortname":"EC","name":"European Commission","jurisdiction":{"code":"EU","label":"European Union"},"pid":null},"level0":{"id":"ec__________::EC::HE","description":"Horizon Europe Framework Programme","name":"HE"},"level1":{"id":"ec__________::EC::HE::HORIZON-TMA-MSCA-SE","description":"HORIZON TMA MSCA Staff Exchanges","name":"HORIZON-TMA-MSCA-SE"},"level2":{"id":null,"description":null,"name":null}},"startDate":"2024-01-01","endDate":"2027-12-31"},{"header":{"relationType":"resultProject","relationClass":"isProducedBy","relatedIdentifier":"corda_____he::7868501db8c07a797fd0de7817d6a54b","relatedRecordType":"project","relationProvenance":"sysimport:crosswalk:repository","trust":"0.9"},"collectedfrom":[{"dsId":"openaire____::3f264f93cf3b0cfc4ede188a6300455c","dsName":"CORDA - COmmon Research DAta Warehouse - Horizon Europe"}],"projectTitle":"Cloud-based Platform-agnostic Adversarial aI Defence framework– CPAID","code":"101168407","funding":{"funder":{"id":"ec__________::EC","shortname":"EC","name":"European Commission","jurisdiction":{"code":"EU","label":"European Union"},"pid":null},"level0":{"id":"ec__________::EC::HE","description":"Horizon Europe Framework Programme","name":"HE"},"level1":{"id":"ec__________::EC::HE::HORIZON-RIA","description":"HORIZON  Research and Innovation Actions","name":"HORIZON-RIA"},"level2":{"id":null,"description":null,"name":null}},"startDate":"2024-10-01","endDate":"2027-09-30"}],"otherTitles":null,"inDiamondJournal":false,"green":true,"isGreen":true,"isInDiamondJournal":false}