[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-117116-en":3,"doc-seo-117116-105":30,"detail-sidebar-cat-0-en-105":91},{"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":4,"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},117116,4398048949847,"Eliana","https://ap-avatar.wpscdn.com/avatar/400002536579ef2da7f?_k=1778318612642679267",8,"Research & Report","Towards Trustworthy Machine Learning with Kernels - Doctor of Philosophy Thesis","Machine Learning has become an indispensable aspect of safety-critical domains such as healthcare, law, and automotive, making trustworthy operation and user trust essential. This thesis improves safety and transparency through principled uncertainty quantification and more effective explainability tools. Kernel Mean Embeddings and Gaussian Processes are developed and used to represent probability distributions with minimal assumptions and to capture uncertainty. Chapters apply them to statistical downscaling with a Deconditional Gaussian process, causal data fusion via combined causal graphs, and explanation methods including RKHS-SHAP and Pref-SHAP.","Towards Trustworthy Machine Learning with Kernels  \nSiu Lun Chau  \nSt.Peter’s College  \nUniversity of Oxford  \nA thesis presented for the degree of Doctor of Philosophy  \nTrinity 2023  \nTo myself, friends, and family.  \nCopyright © 2023 by Siu Lun Chau All Rights Reserved  \nAcknowledgements  \nFirst and foremost, I would like to express my deepest appreciation to my supervisor Professor Dino Sejdinovic for his continuous guidance and encouragements throughout my DPhil, without which I would not have been able to complete this journey. His wisdom and attitude towards both research and life have inspired me to become a better academic and a better individual.  \nI have also had the privilege of working with several incredible researchers and mentors over the past four years. They have all impacted my thinking and approach to research, and I am eternally grateful for their contributions. I extend my sincere appreciation to Dr. Thibaut Lienart for introducing me to Machine Learning in 2016 and supporting my research and professional development since then. Without his contribution, my life would be significantly different. Many thanks to Professor Mihai Cucuringu for his patience and support, especially during the early stages of my DPhil. I am also grateful to Professor Xiaowen Dong, who has always been caring and kind to me. Thanks are also due to Dr. Javier Gonzalez, who had confidence in me during the most challenging period of my DPhil. Lastly, I cannot express my gratitude enough to Dr. Krikamol Muandet. Our discussion on what makes a piece of research meaningful during my stay at the MPI has changed my way of approaching machine learning research. I look forward to continuing working with you all in the future.  \nAside from my mentors, I would also like to thank the EPSRC and the MRC for their generous support of my research. Also, thank you to the Department of Statistics, where I have stayed since I was an undergraduate, and where I always found refuge when I felt lost and doubtful in the past eight years. Thank you to all the staff and researchers who have made this possible.  \nThis endeavour would not have been possible without the support of my lovely colleagues as well. Huge thanks to Bobby He, Deborah Sulem, Edwin Fong, Fan Wu, Henry Kenlay, Jake Fawkes, Leon Law, Lorenzo Pacchiadrdi, Lucian Chan, Michael Zhu Li, Stacy Pu, Veit Wild, William Thomas, and Yin Cong Zhi. Special thanks to Shahine Bouabid, Jean-Francois Ton, and Robert Hu, whom I am lucky enough to have collaborated with and learnt so much from. Their moral support is what kept me sane during the toughest times.  \nI would be in remiss in not mentioning my friends and family, who have supported my journey from the beginning. To Alan Chan, Angel Wong, Bryan Ng, Charig Yang, Cherie Wong, Churchill Ngai, Dominic Li, Gloria Ma, Kenneth Lee, Nicole Hui, Nicholas Yung, Terence Tsui, and Tooki Chu, thank you all for making my Oxford life less nerdy and more fun. To my partner Yidan Gao, a big thank you for bringing joy and new perspectives to my life always. I wish you will have as much fun in your DPhil as I had in mine. To my Grandma: Thank you for raising and feeding me very well all these years. To my sister:  \nThank you for being there whenever I need you. You have always been my biggest fan and soul mate, and I couldn’t be where I am today without your sacrifices and love. To mum: Thank you for your faith in whatever decision I make to pursue my dream. Your wise words are today still my motto today. To dad: Thank you for your tough love, generous support, and understanding all these years. Hope you all are proud of me and will keep watching over me in the coming years.  \nDeclaration  \nI hereby declare that except where specific reference is made to the work of others, the intellectual contents of this dissertation are original and have not been submitted in whole or in part for consideration for anyother degree or qualification. The particulars of my personal","cbCainwB91MZqRRA","https://ap.wps.com/l/cbCainwB91MZqRRA","pdf",16068927,1,213,"English","en",105,"# 1 Introduct","[{\"question\":\"What problem does the thesis address in safety-critical machine learning?\",\"answer\":\"It addresses how to ensure machine learning models function appropriately and build trust among users in safety-critical industries.\"},{\"question\":\"Which techniques are central to the proposed work?\",\"answer\":\"Kernel Mean Embeddings (KME) and Gaussian Processes (GP) are used to represent probability distributions and capture uncertainty effectively.\"},{\"question\":\"How does the thesis approach explainability for kernel methods?\",\"answer\":\"It introduces RKHS-SHAP, which models explainability using Shapley values and estimates the cooperative game value using KMEs, along with a Shapley regulariser.\"}]","Towards Trustworthy Machine Learning with Kernels - Doctor of Philosophy Thesis | PDF",1785673971,537,{"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":86,"head_meta":88,"extra_data":90,"updated_unix":28},"towards-trustworthy-machine-learning-with-kernels-doctor-of-philosophy-thesis","",{"@graph":36,"@context":85},[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/towards-trustworthy-machine-learning-with-kernels-doctor-of-philosophy-thesis/117116/",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","2026-08-02",true,{"@type":65,"interactionType":66,"userInteractionCount":4},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"What problem does the thesis address in safety-critical machine learning?","Question",{"text":75,"@type":76},"It addresses how to ensure machine learning models function appropriately and build trust among users in safety-critical industries.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"Which techniques are central to the proposed work?",{"text":80,"@type":76},"Kernel Mean Embeddings (KME) and Gaussian Processes (GP) are used to represent probability distributions and capture uncertainty effectively.",{"name":82,"@type":73,"acceptedAnswer":83},"How does the thesis approach explainability for kernel methods?",{"text":84,"@type":76},"It introduces RKHS-SHAP, which models explainability using Shapley values and estimates the cooperative game value using KMEs, along with a Shapley regulariser.","https://schema.org",{"og:url":52,"og:type":87,"og:title":13,"og:site_name":59,"og:description":14},"article",{"robots":89,"canonical":52},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,120,123,128,131,135],{"id":20,"doc_module":4,"doc_module_name":46,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":53,"doc_module":4,"doc_module_name":46,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":46,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":46,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":116,"doc_module":4,"doc_module_name":46,"category_name":117,"show_sort_weight":118,"slug":119},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":121,"slug":122},30,"research-report",{"id":124,"doc_module":4,"doc_module_name":46,"category_name":125,"show_sort_weight":126,"slug":127},9,"Religion & Spirituality",20,"religion-spirituality",{"id":126,"doc_module":4,"doc_module_name":46,"category_name":129,"show_sort_weight":126,"slug":130},"World Cup","world-cup",{"id":132,"doc_module":4,"doc_module_name":46,"category_name":133,"show_sort_weight":132,"slug":134},10,"Lifestyle","lifestyle",{"id":136,"doc_module":4,"doc_module_name":46,"category_name":137,"show_sort_weight":106,"slug":138},19,"General","general"]