[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-125838-en":3,"doc-seo-125838-105":31,"detail-sidebar-cat-0-en-105":93},{"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":21,"is_downloadable":21,"audit_status":21,"page_count":22,"language":23,"language_code":24,"site_id":25,"html_lang":24,"table_of_contents":26,"faqs":27,"seo_title":28,"seo_description":14,"update_tm":29,"read_time":30},125838,1099523885074,"Ivy","https://ap-avatar.wpscdn.com/davatar_9964176cb1d06d4a9deccf72a44ae3dc",8,"Research & Report","Enhancing Robustness of Machine Learning Models Against Adversarial Attacks","Machine learning models are essential across many applications, yet they are increasingly exposed to adversarial attacks that introduce subtle input changes to mislead predictions. This thesis focuses on strengthening model dependability and security by addressing limitations in prior work, which often evaluates narrow attack types and restricted algorithm sets. A two-phase approach simulates attacks such as BIM, DeepFool, and FGSM on models trained with MNIST and CIFAR-10, then analyzes defense strategies that reduce sensitivity to input perturbations, improving resilience for future secure ML systems.","Portland State University  \nPDXScholar  \n\n| University Honors Theses | University Honors College |\n| --- | --- |\n| Spring 6-2024\u003Cbr>Enhancing Robustness of Machine Learning Models Against Adversarial Attacks\u003Cbr>Ronak Guliani\u003Cbr>Portland State University\u003Cbr>Follow this and additional works at: [https://pdxscholar.library.pdx.edu/honorstheses](https://pdxscholar.library.pdx.edu/honorstheses)\u003Cbr> Part of the Artificial Intelligence and Robotics Commons, and the Theory and Algorithms Commons Let us know how access to this document benefits you. |  |\n\nRecommended Citation  \nGuliani, Ronak, \"Enhancing Robustness of Machine Learning Models Against Adversarial Attacks\" (2024) . University Honors Theses. Paper 1549.  \n[https://doi.org/10.15760/honors.1581](https://doi.org/10.15760/honors.1581)  \nThis Thesis is brought to you for free and open access. It has been accepted for inclusion in University Honors Theses by an authorized administrator of PDXScholar. Please contact us if we can make this document more accessible: [pdxscholar@pdx.edu](pdxscholar@pdx.edu).  \nEnhancing Robustness of Machine Learning Models against Adversarial Attacks  \nby  \nRonak Guliani  \nAn undergraduate honors thesis submitted in partial fulfillment of the requirements for the degree of  \nBachelor of Science  \nin  \nUniversity Honors  \nand  \nComputer Science  \nThesis Advisor  \nNirupama Bulusu  \nPortland State University  \n2024  \nABSTRACT  \nMachine learning models are integral for numerous applications, but they remain increasingly vulnerable to adversarial attacks. These attacks involve subtle manipulation of input data to deceive models, presenting a critical threat to their dependability and security. This thesis addresses the need for strengthening these models against such adversarial attacks. Prior research has primarily focused on identifying specific types of adversarial attacks on a limited range of ML algorithms. However, there is a gap in the evaluation of model resilience across algorithms and in the development of effective defense mechanisms. To bridge this gap, this work adopts a two-phase approach. First, it simulates attacks like the Basic Iterative Method (BIM), DeepFool, and Fast Gradient Sign Method (FGSM)on common ML models trained on MNIST and CIFAR-10, which are used for image processing. This thesis will then discuss defensive strategies, which reduces the sensitivity of the model to input changes to improve model resilience against attacks. The findings are aimed to benefit ML researchers to develop more secure and robust ML systems.  \nTABLE OF CONTENTS  \nABSTRACT i  \n1 INTRODUCTION 1  \n2 LITERATURE REVIEW 2  \n2.1 Applications 2  \n2.1.1 Natural Language Processing 2  \n2.1.2 Healthcare 3  \n2.1.3 Autonomous Systems 3  \n2.1.4 Finance 4  \n2.2 Emergence of Adversarial Vulnerabilities 4  \n2.3 Advancements in Generating Adversarial Examples 5  \n2.4 Developing Defensive Mechanisms 5  \n2.5 Synthesis of Adversarial Research 6  \n3 METHODOLOGY 8  \n3.1 Conceptual Framework and Experimental Design 8  \n3.2 Adversarial Attack Simulations 9  \n3.3 Defensive Strategies Evaluation 9  \n3.4 Implementation and Training Procedures 10  \n4 FINDINGSAND DISCUSSION 18  \n4.1 Baseline Model Training and Evaluation 18  \n4.2 Evaluation of Adversarial Attacks 20  \n4.3 Evaluation of Defensive Strategies 25  \n5 DISCUSSION 28  \n5. 1 Implications for Future Research 28  \n5.2 Limitations of the Study 28  \nReferences 30  \n1 INTRODUCTION  \nFrom speech and image recognition to predictive tools in healthcare, machine learning models (ML) are crucial in various applications . As these models become essential in both critical infrastructures and personal devices, ensuring their security and dependability are increasingly important. A major threat to their integrity are adversarial attacks—sophisticated techniques designed to deceive ML systems by manipulating input data.  \nExisting research on adversarial attacks reveals that even the most advanced ML models are susceptible to these ","cbCaio5BPYCSHyYG","https://ap.wps.com/l/cbCaio5BPYCSHyYG","pdf",2053045,10,1,35,"English","en",105,"# Abstract\n# 1 Introduction\n# 2 Literature Review\n## 2.1 Applications\n## 2.2 Emergence of Adversarial Vulnerabilities\n## 2.3 Advancements in Generating Adversarial Examples\n## 2.4 Developing Defensive Mechanisms\n## 2.5 Synthesis of Adversarial Research\n# 3 Methodology\n## 3.1 Conceptual Framework and Experimental Design\n## 3.2 Adversarial Attack Simulations\n## 3.3 Defensive Strategies Evaluation\n## 3.4 Implementation and Training Procedures\n# 4 Findings and Discussion\n## 4.1 Baseline Model Training and Evaluation\n## 4.2 Evaluation of Adversarial Attacks\n## 4.3 Evaluation of Defensive Strategies","[{\"question\":\"What problem does the thesis address?\",\"answer\":\"It addresses the vulnerability of machine learning models to adversarial attacks, where subtle input manipulations can cause incorrect predictions and undermine reliability.\"},{\"question\":\"Which adversarial attacks are used in the experiments?\",\"answer\":\"The thesis simulates Basic Iterative Method (BIM), DeepFool, and Fast Gradient Sign Method (FGSM) on common machine learning models.\"},{\"question\":\"How does the thesis evaluate defenses?\",\"answer\":\"It discusses defensive strategies aimed at reducing model sensitivity to input changes, then evaluates how these defenses improve resilience against adversarial perturbations.\"}]","Enhancing Robustness of Machine Learning Models Against Adversarial Attacks | PDF",1785901511,88,{"code":4,"msg":32,"data":33},"ok",{"site_id":25,"language":24,"slug":34,"title":13,"keywords":35,"description":14,"schema_data":36,"social_meta":88,"head_meta":90,"extra_data":92,"updated_unix":29},"enhancing-robustness-of-machine-learning-models-against-adversarial-attacks","",{"@graph":37,"@context":87},[38,55,70],{"@type":39,"itemListElement":40},"BreadcrumbList",[41,45,49,52],{"item":42,"name":43,"@type":44,"position":21},"https://docshare.wps.com","Home","ListItem",{"item":46,"name":47,"@type":44,"position":48},"https://docshare.wps.com/document/","Document",2,{"item":50,"name":12,"@type":44,"position":51},"https://docshare.wps.com/document/research-report/",3,{"item":53,"name":13,"@type":44,"position":54},"https://docshare.wps.com/document/enhancing-robustness-of-machine-learning-models-against-adversarial-attacks/125838/",4,{"url":53,"name":13,"@type":56,"author":57,"headline":13,"publisher":59,"fileFormat":62,"inLanguage":24,"description":14,"dateModified":63,"datePublished":64,"encodingFormat":62,"isAccessibleForFree":65,"interactionStatistic":66},"DigitalDocument",{"name":9,"@type":58},"Person",{"url":42,"name":60,"@type":61},"DocShare","Organization","application/pdf","2026-08-23","2026-08-05",true,{"@type":67,"interactionType":68,"userInteractionCount":20},"InteractionCounter",{"@type":69},"ViewAction",{"@type":71,"mainEntity":72},"FAQPage",[73,79,83],{"name":74,"@type":75,"acceptedAnswer":76},"What problem does the thesis address?","Question",{"text":77,"@type":78},"It addresses the vulnerability of machine learning models to adversarial attacks, where subtle input manipulations can cause incorrect predictions and undermine reliability.","Answer",{"name":80,"@type":75,"acceptedAnswer":81},"Which adversarial attacks are used in the experiments?",{"text":82,"@type":78},"The thesis simulates Basic Iterative Method (BIM), DeepFool, and Fast Gradient Sign Method (FGSM) on common machine learning models.",{"name":84,"@type":75,"acceptedAnswer":85},"How does the thesis evaluate defenses?",{"text":86,"@type":78},"It discusses defensive strategies aimed at reducing model sensitivity to input changes, then evaluates how these defenses improve resilience against adversarial perturbations.","https://schema.org",{"og:url":53,"og:type":89,"og:title":13,"og:site_name":60,"og:description":14},"article",{"robots":91,"canonical":53},"index,follow",{"doc_id":7,"site_id":25},{"code":4,"msg":5,"data":94},[95,99,103,107,112,117,122,125,130,133,136],{"id":21,"doc_module":4,"doc_module_name":47,"category_name":96,"show_sort_weight":97,"slug":98},"Story & Novel",90,"story-novel",{"id":48,"doc_module":4,"doc_module_name":47,"category_name":100,"show_sort_weight":101,"slug":102},"Literature",80,"literature",{"id":54,"doc_module":4,"doc_module_name":47,"category_name":104,"show_sort_weight":105,"slug":106},"Exam",70,"exam",{"id":108,"doc_module":4,"doc_module_name":47,"category_name":109,"show_sort_weight":110,"slug":111},5,"Comic",60,"comic",{"id":113,"doc_module":4,"doc_module_name":47,"category_name":114,"show_sort_weight":115,"slug":116},6,"Technology",50,"technology",{"id":118,"doc_module":4,"doc_module_name":47,"category_name":119,"show_sort_weight":120,"slug":121},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":47,"category_name":12,"show_sort_weight":123,"slug":124},30,"research-report",{"id":126,"doc_module":4,"doc_module_name":47,"category_name":127,"show_sort_weight":128,"slug":129},9,"Religion & Spirituality",20,"religion-spirituality",{"id":128,"doc_module":4,"doc_module_name":47,"category_name":131,"show_sort_weight":128,"slug":132},"World Cup","world-cup",{"id":20,"doc_module":4,"doc_module_name":47,"category_name":134,"show_sort_weight":20,"slug":135},"Lifestyle","lifestyle",{"id":137,"doc_module":4,"doc_module_name":47,"category_name":138,"show_sort_weight":108,"slug":139},19,"General","general"]