[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-121072-en":3,"doc-seo-121072-105":30,"detail-sidebar-cat-0-en-105":95},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":20,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":21,"language":22,"language_code":23,"site_id":24,"html_lang":23,"table_of_contents":25,"faqs":26,"seo_title":27,"seo_description":14,"update_tm":28,"read_time":29},121072,8796095462418,"Noah","https://ap-avatar.wpscdn.com/avatar/80000253c1241d02b47?x-image-process=image/resize,m_fixed,w_180,h_180&k=1778826106357471780",8,"Research & Report","Adversarial Machine Learning in IoT - Vulnerability Analysis and Robustness - PhD Thesis","Machine learning models used in Internet of Things (IoT) applications—such as traffic profiling, network security, and device identification—are susceptible to adversarial attacks that can cause misclassifications and operational disruption. This thesis investigates the security of ML models in IoT settings by developing an evaluation methodology for white- and black-box attacks and generating imperceptible adversarial examples. Results show IoT device classifiers remain vulnerable, motivating new defenses. The work further introduces a discretization-based ensemble robustness approach, validating improved resilience against multiple adversarial attacks, including deep intrusion detection models in Edge IIoT with datasets reflecting realistic scenarios.","Adversarial Machine Learning in IoT: Vulnerability Analysis and Roboustness  \nAuthor:  \nNamvar, Anahita  \nPublication Date:  \n2024  \nDOI:  \n[https://doi.org/10.26190/unsworks/30278](https://doi.org/10.26190/unsworks/30278)  \nLicense:  \n[https://creativecommons.org/licenses/by/4.0/](https://creativecommons.org/licenses/by/4.0/)  \nLink to license to see what you are allowed to do with this resource.  \nDownloaded from [http://hdl.handle. net/1959.4/102465](http://hdl.handle. net/1959.4/102465) in [https://](https://)[ ](https://)[unsworks. unsw.edu.au](unsworks. unsw.edu.au) on 2024-10-11  \nAdversarial Machine Learning in IoT: Vulnerability Analysis and Robustness  \nAnahita Namvar  \nA thesis in fulfilment of the requirements for the degree of  \nDoctor of Philosophy  \nSchool of Computer Science and Engineering Faculty of Engineering The University of New South Wales  \nOctober 2023  \nTHE UNIVERSITY OF NEW SOUTH WALES  \nThesis/Dissertation Sheet  \nSurname or Family name: Namvar  \nFirst name: Anahita Other name/s:  \nAbbreviation for degree as given in the University calendar: PhD  \nSchool: School of Computer Science and Engineering Faculty: Faculty of Engineering  \nTitle: Adversarial Machine Learning in IoT: Vulnerability Analysis and Robustness  \nAbstract  \nMachine learning (ML) has become widely used in various Internet of Things (IoT) applications, including tasks like traffic profiling, network security, and identifying IoT devices. However, it is essential to acknowledge that ML models are vulnerable to adversarial attacks, which can lead to misclassifications and disruptions to the specific application contexts in which they are being used. While adversarial attacks have been extensively studied in fields such as computer vision, there is still ample opportunity for further research and improvement when applied within the context of IoT.  \nThis thesis presents a comprehensive investigation into the security of ML models in the IoT landscape . It offers novel insights and defense strategies to enhance the resilience of ML-powered IoT systems against adversarial threats. To begin, we develop a research methodology for evaluating machine learning models’ vulnerability to white and black-box attacks. We then create imperceptible adversarial examples using state-of-the-art attack techniques. Our investigation focuses on the IoT device identification case study, demonstrating the effectiveness of our methodology. The findings underscore the susceptibility of IoT device classifiers to adversarial attacks, emphasizing the urgent need for enhanced security measures within this domain.  \nSecond, we propose a novel discretization-based ensemble robustness methodology. This mechanism leverages discretization techniques and ensemble learning to fortify models against adversarial manipulations, ensuring reliable and secure performance . To evaluate the effectiveness of the proposed models, we present the enhanced security of IoT device identification models against previously generated adversarial attacks. In addition, we validate the efficacy of discretization methods as defense mechanisms by showcasing their ability to fortify IoT device identification classifiers against a spectrum of adversarial attacks.  \nWe expand our research scope to explore the Edge Industrial Internet of Things(IIoT) . Here, we apply our proposed research methodology and defense mechanism to evaluate and enhance the security of machine learning models within a real-world Edge IIoT case study, utilizing a recently published Edge-IIoT dataset designed to resemble real-world scenarios. Our investigation extends to deep learning models tailored for intrusion detection, encompassing models with various layers and complexities, including the intricate Resnet50 deep model. We systematically generate imperceptible adversarial inputs to these models, investigating the vulnerability of deep intrusion models to adversarial attacks in the IIoT domain. To combat these vulnerabi","cbCaivet0nRP9kRb","https://ap.wps.com/l/cbCaivet0nRP9kRb","pdf",7421946,1,157,"English","en",105,"# Abstract\n## Threat model and evaluation methodology\n## Adversarial examples for IoT device identification\n## Discretization-based ensemble robustness\n## Extension to Edge IIoT and intrusion detection\n## Conclusions and contributions","[{\"question\":\"What problem does the thesis address in IoT machine learning security?\",\"answer\":\"It addresses how ML models deployed in IoT can be compromised by adversarial attacks, leading to misclassification and disruption of application-specific tasks.\"},{\"question\":\"How does the thesis evaluate vulnerability to adversarial attacks?\",\"answer\":\"It develops a methodology to assess ML models under white-box and black-box attack settings, then generates imperceptible adversarial examples.\"},{\"question\":\"What defense approach does the thesis propose and what does it improve?\",\"answer\":\"It proposes a discretization-based ensemble robustness method that leverages discretization and ensemble learning to strengthen model resilience against adversarial manipulations.\"},{\"question\":\"How is the research extended beyond basic IoT device identification?\",\"answer\":\"It applies the methodology and defense mechanism to Edge IIoT, including deep intrusion detection models such as Resnet50, using an Edge-IIoT dataset designed to resemble real-world scenarios.\"}]","Adversarial Machine Learning in IoT - Vulnerability Analysis and Robustness - PhD Thesis | PDF",1785733594,396,{"code":4,"msg":31,"data":32},"ok",{"site_id":24,"language":23,"slug":33,"title":13,"keywords":34,"description":14,"schema_data":35,"social_meta":90,"head_meta":92,"extra_data":94,"updated_unix":28},"adversarial-machine-learning-in-iot-vulnerability-analysis-and-robustness-phd-thesis","",{"@graph":36,"@context":89},[37,54,68],{"@type":38,"itemListElement":39},"BreadcrumbList",[40,44,48,51],{"item":41,"name":42,"@type":43,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":45,"name":46,"@type":43,"position":47},"https://docshare.wps.com/document/","Document",2,{"item":49,"name":12,"@type":43,"position":50},"https://docshare.wps.com/document/research-report/",3,{"item":52,"name":13,"@type":43,"position":53},"https://docshare.wps.com/document/adversarial-machine-learning-in-iot-vulnerability-analysis-and-robustness-phd-thesis/121072/",4,{"url":52,"name":13,"@type":55,"author":56,"headline":13,"publisher":58,"fileFormat":61,"inLanguage":23,"description":14,"dateModified":62,"datePublished":62,"encodingFormat":61,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":57},"Person",{"url":41,"name":59,"@type":60},"DocShare","Organization","application/pdf","2026-08-03",true,{"@type":65,"interactionType":66,"userInteractionCount":20},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81,85],{"name":72,"@type":73,"acceptedAnswer":74},"What problem does the thesis address in IoT machine learning security?","Question",{"text":75,"@type":76},"It addresses how ML models deployed in IoT can be compromised by adversarial attacks, leading to misclassification and disruption of application-specific tasks.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How does the thesis evaluate vulnerability to adversarial attacks?",{"text":80,"@type":76},"It develops a methodology to assess ML models under white-box and black-box attack settings, then generates imperceptible adversarial examples.",{"name":82,"@type":73,"acceptedAnswer":83},"What defense approach does the thesis propose and what does it improve?",{"text":84,"@type":76},"It proposes a discretization-based ensemble robustness method that leverages discretization and ensemble learning to strengthen model resilience against adversarial manipulations.",{"name":86,"@type":73,"acceptedAnswer":87},"How is the research extended beyond basic IoT device identification?",{"text":88,"@type":76},"It applies the methodology and defense mechanism to Edge IIoT, including deep intrusion detection models such as Resnet50, using an Edge-IIoT dataset designed to resemble real-world scenarios.","https://schema.org",{"og:url":52,"og:type":91,"og:title":13,"og:site_name":59,"og:description":14},"article",{"robots":93,"canonical":52},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":96},[97,101,105,109,114,119,124,127,132,135,139],{"id":20,"doc_module":4,"doc_module_name":46,"category_name":98,"show_sort_weight":99,"slug":100},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":102,"show_sort_weight":103,"slug":104},"Literature",80,"literature",{"id":53,"doc_module":4,"doc_module_name":46,"category_name":106,"show_sort_weight":107,"slug":108},"Exam",70,"exam",{"id":110,"doc_module":4,"doc_module_name":46,"category_name":111,"show_sort_weight":112,"slug":113},5,"Comic",60,"comic",{"id":115,"doc_module":4,"doc_module_name":46,"category_name":116,"show_sort_weight":117,"slug":118},6,"Technology",50,"technology",{"id":120,"doc_module":4,"doc_module_name":46,"category_name":121,"show_sort_weight":122,"slug":123},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":125,"slug":126},30,"research-report",{"id":128,"doc_module":4,"doc_module_name":46,"category_name":129,"show_sort_weight":130,"slug":131},9,"Religion & Spirituality",20,"religion-spirituality",{"id":130,"doc_module":4,"doc_module_name":46,"category_name":133,"show_sort_weight":130,"slug":134},"World Cup","world-cup",{"id":136,"doc_module":4,"doc_module_name":46,"category_name":137,"show_sort_weight":136,"slug":138},10,"Lifestyle","lifestyle",{"id":140,"doc_module":4,"doc_module_name":46,"category_name":141,"show_sort_weight":110,"slug":142},19,"General","general"]