[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-seo-290549-105":3,"detail-sidebar-cat-0-en-105":81,"doc-detail-290549-en":131},{"code":4,"msg":5,"data":6},0,"ok",{"site_id":7,"language":8,"slug":9,"title":10,"keywords":11,"description":12,"schema_data":13,"social_meta":74,"head_meta":76,"extra_data":78,"updated_unix":80},105,"en","machine-learning-for-mitigating-adversarial-influence-in-networked-systems-methods-and-applications-in-global-supply-chains","Machine Learning for Mitigating Adversarial Influence in Networked Systems - Methods and Applications in Global Supply-Chains","","Bad actors, including forced labour firms, exploit the conditions created by modern supply networks. As network complexity grows, visibility declines, detection becomes more difficult, and these actors gain wider reach. This thesis argues that machine learning provides a practical route to detect and mitigate supply-chain bad actors. It introduces data characterization and learning-based detection for forced labour, and a graph neural network approach to identify and recommend substitutions for unsustainable food products.",{"@graph":14,"@context":73},[15,34,56],{"@type":16,"itemListElement":17},"BreadcrumbList",[18,23,27,31],{"item":19,"name":20,"@type":21,"position":22},"https://docshare.wps.com","Home","ListItem",1,{"item":24,"name":25,"@type":21,"position":26},"https://docshare.wps.com/document/","Document",2,{"item":28,"name":29,"@type":21,"position":30},"https://docshare.wps.com/document/research-report/","Research & Report",3,{"item":32,"name":10,"@type":21,"position":33},"https://docshare.wps.com/document/machine-learning-for-mitigating-adversarial-influence-in-networked-systems-methods-and-applications-in-global-supply-chains/290549/",4,{"url":32,"name":10,"@type":35,"image":36,"author":41,"headline":10,"publisher":44,"fileFormat":47,"inLanguage":8,"description":12,"dateModified":48,"datePublished":49,"encodingFormat":47,"isAccessibleForFree":50,"interactionStatistic":51},"DigitalDocument",{"url":37,"@type":38,"width":39,"height":40},"https://docshare.wps.com/thumbnails/machine-learning-for-mitigating-adversarial-influence-in-networked-systems-methods-and-applications-in-global-supply-chains/290549.png","ImageObject",300,407,{"name":42,"@type":43},"Noah","Person",{"url":19,"name":45,"@type":46},"DocShare","Organization","application/pdf","2026-10-27","2026-09-17",true,{"@type":52,"interactionType":53,"userInteractionCount":55},"InteractionCounter",{"@type":54},"ViewAction",12,{"@type":57,"mainEntity":58},"FAQPage",[59,65,69],{"name":60,"@type":61,"acceptedAnswer":62},"Why does adversarial influence become harder to mitigate in modern supply networks?","Question",{"text":63,"@type":64},"Growth in network complexity reduces visibility, makes detection harder, and expands the reach of bad actors.","Answer",{"name":66,"@type":61,"acceptedAnswer":67},"How does the thesis use machine learning to address forced labour in global supply networks?",{"text":68,"@type":64},"It presents a novel data study that characterizes the network features of forced labour firms and then leverages machine learning to detect whether forced labour is used and whether it likely exists within a firm’s supply network.",{"name":70,"@type":61,"acceptedAnswer":71},"What additional limitations in existing machine learning methods does the thesis tackle?",{"text":72,"@type":64},"It addresses the mismatch between typical i.i.d assumptions and real dynamic, temporal settings by proposing an optimiser for fast training from non-i.i.d data streams.","https://schema.org",{"og:url":32,"og:type":75,"og:title":10,"og:site_name":45,"og:description":12},"article",{"robots":77,"canonical":32},"index,follow",{"doc_id":79,"site_id":7},290549,1789642038,{"code":4,"msg":82,"data":83},"success",[84,88,92,96,101,106,111,115,120,123,127],{"id":22,"doc_module":4,"doc_module_name":25,"category_name":85,"show_sort_weight":86,"slug":87},"Story & Novel",90,"story-novel",{"id":26,"doc_module":4,"doc_module_name":25,"category_name":89,"show_sort_weight":90,"slug":91},"Literature",80,"literature",{"id":33,"doc_module":4,"doc_module_name":25,"category_name":93,"show_sort_weight":94,"slug":95},"Exam",70,"exam",{"id":97,"doc_module":4,"doc_module_name":25,"category_name":98,"show_sort_weight":99,"slug":100},5,"Comic",60,"comic",{"id":102,"doc_module":4,"doc_module_name":25,"category_name":103,"show_sort_weight":104,"slug":105},6,"Technology",50,"technology",{"id":107,"doc_module":4,"doc_module_name":25,"category_name":108,"show_sort_weight":109,"slug":110},7,"Healthcare",40,"healthcare",{"id":112,"doc_module":4,"doc_module_name":25,"category_name":29,"show_sort_weight":113,"slug":114},8,30,"research-report",{"id":116,"doc_module":4,"doc_module_name":25,"category_name":117,"show_sort_weight":118,"slug":119},9,"Religion & Spirituality",20,"religion-spirituality",{"id":118,"doc_module":4,"doc_module_name":25,"category_name":121,"show_sort_weight":118,"slug":122},"World Cup","world-cup",{"id":124,"doc_module":4,"doc_module_name":25,"category_name":125,"show_sort_weight":124,"slug":126},10,"Lifestyle","lifestyle",{"id":128,"doc_module":4,"doc_module_name":25,"category_name":129,"show_sort_weight":97,"slug":130},19,"General","general",{"code":4,"msg":82,"data":132},{"doc_id":79,"user_id":133,"nickname":42,"user_avatar":134,"doc_module":4,"category_id":112,"category_name":29,"doc_title":10,"doc_description":12,"doc_content":135,"file_id":136,"file_url":137,"file_type":138,"file_size":139,"view_count":55,"is_deleted":4,"is_public":22,"is_downloadable":22,"audit_status":22,"page_count":140,"language":141,"language_code":8,"site_id":7,"html_lang":8,"table_of_contents":142,"faqs":143,"seo_title":144,"seo_description":12,"update_tm":80,"read_time":145,"preview_page_count":4},8796095462418,"https://ap-avatar.wpscdn.com/avatar/80000253c1241d02b47?x-image-process=image/resize,m_fixed,w_180,h_180&k=1778826106357471780","Machine Learning for Mitigating Adversarial Influence in Networked  \nSystems  \nMethods and Applications in Global Supply-Chains  \nJack William Foster  \nUniversity of Cambridge  \nThis thesis is submitted for the degree of Doctor of Philosophy  \nCorpus Christi College November 2025  \nDeclaration  \nThis thesis is the result of my own work and includes nothing which is the outcome of work done in collaboration or content generated by artificial intelligence, except as declared in the preface and specified in the text. It is not substantially the same as any work that has already been submitted, or is being concurrently submitted, for any degree, diploma or other qualification at the University of Cambridge or any other University or similar institution except as declared in the preface and specified in the text. It does not exceed the prescribed word limit for the relevant Degree Committee.  \nAbstract  \nBad actors, such as forced labour firms, thrive in the conditions created by modern supplynetworks. Growth in network complexity reduces visibility, makes detection harder, and expands the reach of bad actors. Mitigating their influence is vital to the existence of ethical, robust, and trustworthy supply-networks. This thesis argues that machine learning is a practical path to both detect and mitigate supply-chain bad actors. The first contributions of this thesis belong to the supply-chain literature, starting with a novel data-study of forced labour in global supply-networks, where the network features of forced labour firms are first characterised, and then leveraged in conjunction with machine learning to help detect whether an organisation uses forced labour, and whether forced labour likely exists within a firm’s supply-network. Next, a case-study on unsustainable food products in the food supply-chain demonstrates not only that machine learning can detect bad actors, but that it can also offer actionable insights to buyers, as a graph neural network is leveraged to recommend substitutions to allow the removal of unsustainable products.  \nSubsequent chapters address key limitations in existing machine learning methods. Supervised machine learning is typically predicated upon the assumption that data is independent and identically distributed (i.i.d), an assumption that does not hold for real-world settings that are dynamic and inherently temporal. As such, the first contribution to the field of machine learning is a novel optimiser that supports fast and efficient model training from non-i.i.d data streams. Finally, we make a series of significant contributions to the emerging field of machine unlearning, which facilitates a model’s ability to forget harmful, outdated, or private information acquired in training, without damaging model utility on the wider data.  \nTogether, this thesis presents a blend of real-world insights and methods for specific supply-chain bad actor challenges, alongside a complimentary set of contributions that alleviate key challenges faced by machine learning models when deployed in real-world settings. We evaluate on curated data in specific domains, and we scope our claims accordingly. Further work is needed to demonstrate generalisability to the full set of supply-chain bad actors, and to operationalise our contributions in real industry settings.  \nAcknowledgements  \nIt is no coincidence that my best work has arisen from my periods of greatest collaboration. And so to my friends and peers, Stefan, Kyle, and Elijah, I extend my warmest thanks foryour support and inspiration over these past years. To my lab mates, I extend you the same gratitude.  \nI thank my supervisor, Prof. Alexandra Brintrup, for her support, guidance, and understanding throughout the PhD. I would also like to thank my advisor, Prof. Ajith Parlikad. It has been a pleasure being your student.  \nTo the EPSRC AgriFoRwArdS CDT, I thank you for providing me with this opportunity. It has been a privilege to undertake this PhD, and fo","cbCaicem7r2sWzu8","https://ap.wps.com/l/cbCaicem7r2sWzu8","pdf",5617330,210,"English","# Table of contents\n## List of figures\n## List of tables\n## 1 Introduction\n## 1.1 Bad Actors in Networked Systems\n## 1.2 The Role and Limitations of Machine Learning\n## 1.3 Research Gaps\n## 1.4 Research Aims, Questions, and Objectives\n## 1.5 Research Framework\n## 1.6 Research Scope\n## 1.7 Research Methodology\n## 1.8 Contributions\n## 1.9 Thesis Structure\n## 1.10 Publications\n## 2 Literature Review\n## 2.1 Supply-Chain Bad Actors","[{\"question\":\"Why does adversarial influence become harder to mitigate in modern supply networks?\",\"answer\":\"Growth in network complexity reduces visibility, makes detection harder, and expands the reach of bad actors.\"},{\"question\":\"How does the thesis use machine learning to address forced labour in global supply networks?\",\"answer\":\"It presents a novel data study that characterizes the network features of forced labour firms and then leverages machine learning to detect whether forced labour is used and whether it likely exists within a firm’s supply network.\"},{\"question\":\"What additional limitations in existing machine learning methods does the thesis tackle?\",\"answer\":\"It addresses the mismatch between typical i.i.d assumptions and real dynamic, temporal settings by proposing an optimiser for fast training from non-i.i.d data streams.\"}]","Machine Learning for Mitigating Adversarial Influence in Networked Systems - Methods and Applications in Global Supply-Chains | PDF",529]