[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-119979-en":3,"doc-seo-119979-105":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},119979,549758252649,"Ivy","https://ap-avatar.wpscdn.com/avatar/8000253669c5317157?_k=1778319167496531819",8,"Research & Report","Deep Generative Models for Biology - Represent, Predict, Design","Deep generative models reshape artificial intelligence by enabling creation of novel objects that imitate or extrapolate from training data, and by changing how texts, images, speech, and programs are represented and consumed. Their scientific impact extends beyond AI to domains such as mathematical problem solving and fast, accurate simulations in high-energy physics, as well as rapid weather forecasting. In computational biology, they promise better understanding of complex biological processes, improved design of drugs and therapies, and forecasting viral evolution, while facing challenges from massive spaces, multimodal data, and mixed structured–unstructured components. This thesis develops generative modeling frameworks for these questions.","Deep Generative Models for Biology: Represent, Predict, Design  \nPascal Notin Department of Computer Science  \nUniversity of Oxford  \nA thesis submitted for the degree of Doctor of Philosophy  \n—  \nSt Cross College Trinity 2023  \nAcknowledgements  \nI would like to first express my deepest gratitude to my supervisor, Yarin Gal, who has been an exceptional mentor and thought partner throughout my DPhil. His humility and brightness have been a constant source of inspiration, and Iam forever grateful for his trust when he accepted me into the Oxford Applied and Theoretical Machine Learning group. I look forward to many more years of collaborating together.  \nI am incredibly fortunate to have been part of OATML, and I extend my heartfelt thanks to my wonderful colleagues there, especially those I have had the opportunity to collaborate with: Andrew Jesson, Clare Lyle, Ruben Weitzman, Lood van Niekerk, Shabbir Khan, Lisa Schut, Neil Band, and Jannik Kossen. Special thanks to Aidan Gomez for a fantastic research collaboration on many projects and from whom I have learned a great deal. I wish him all the best as he sets out to revolutionize the way we use and analyze language at Cohere. Cheers to Joost van Amersfoort, Angelos Filos and Panos Tigas-for the laughs, camaraderie, and inspiration-to Seb Farquhar, for showing me that a path in scientific research is possible after years in management consulting, and to Tim Rudner, for his great taste in literature. Thank you to Milad Alizadeh, Joost van Amersfoort, Lewis Smith, Andreas Kirsch, Lorenz Kuhn, and Muhammed Razzak for setting up and maintaining the OAT cloud, which has been an invaluable resource for several projects in this thesis. I would like to also acknowledge the teams behind the ARC and JADE compute clusters at Oxford for their support, which has made much of my research possible. I am indebted to my collaborators at the Marks Lab for teaching me almost everything I know about Computational Biology. Special appreciation goes to Jonny and Mafalda for patiently going through the basics with me when we started working on EVE – I am thankful for our numerous collaborations and wish them the best as they establish their own lab in Barcelona. Additionally, I am grateful to Nikki Thadani, Sarah Gurev, Noor Youssef for teaching me most things I know about viruses; Sam Berry for everything I know about GPCRs; Aaron Kollasch and Dan Ritter for a wonderful collaboration on indels; and Nathan Rollins and Chris Sander for their help in navigating the complex, inspiring and fast-moving field of protein design. Last but not least, I want to express my deepest thanks to Debora Marks, without whom my DPhil journey would have been markedly different. Her  \nmentorship and guidance throughout our many joint projects have been invaluable, and I eagerly anticipate our continued collaboration in the years ahead.  \nI have had the immense pleasure to collaborate with Miguel Lobato, who taught me the basics of molecular optimization and provided invaluable guidance in the early stages of my DPhil. I would like to acknowledge Mark Woolrich and Chetan Gohil for our collaboration on fascinating neuroscience projects.  \nI am deeply grateful to GSK for their support throughout my DPhil, and I extend my appreciation to Lindsay Edwards for settings things up and Kim Branson for his ongoing support thereafter. I am particularly thankful to Patrick Schwab, who has provided guidance and thought leadership on many collaborations – from conducting cutting-edge research projects at the intersection of Active Learning and Biology, to bootstrapping the Machine Learning for Drug Discovery workshop at ICLR, to running the GeneDisco and CausalBench challenges. I would also like to thank Stefan Bauer, Arash Mehrjou, Ashkan Soleymani and Mathieu Chevalley for their partnership during these various initiatives, and would like to acknowledge my numerous collaborators at the Machine Learning for Drug Discovery workshop ","cbCaigD2If4d3YqI","https://ap.wps.com/l/cbCaigD2If4d3YqI","pdf",30177009,1,255,"English","en",105,"# Abstract\n## Thesis motivation and scope\n## Deep generative modeling in computational biology","[{\"question\":\"What are deep generative models expected to enable in this thesis context?\",\"answer\":\"They enable generating novel objects by imitating or extrapolating from training data and support access to information across multiple modalities such as text, images, speech, and programs.\"},{\"question\":\"Which computational biology applications motivate the work?\",\"answer\":\"The thesis highlights improving understanding of complex biological processes, designing new drugs and therapies, and forecasting viral evolution during pandemics.\"},{\"question\":\"What makes biological modeling challenging for generative models?\",\"answer\":\"Biological objects involve massive search spaces, multiple complementary data modalities, and interactions between highly structured and relatively unstructured components.\"}]","Deep Generative Models for Biology - Represent, Predict, Design | PDF",1785727414,643,{"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},"deep-generative-models-for-biology-represent-predict-design","",{"@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/deep-generative-models-for-biology-represent-predict-design/119979/",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":4},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"What are deep generative models expected to enable in this thesis context?","Question",{"text":75,"@type":76},"They enable generating novel objects by imitating or extrapolating from training data and support access to information across multiple modalities such as text, images, speech, and programs.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"Which computational biology applications motivate the work?",{"text":80,"@type":76},"The thesis highlights improving understanding of complex biological processes, designing new drugs and therapies, and forecasting viral evolution during pandemics.",{"name":82,"@type":73,"acceptedAnswer":83},"What makes biological modeling challenging for generative models?",{"text":84,"@type":76},"Biological objects involve massive search spaces, multiple complementary data modalities, and interactions between highly structured and relatively unstructured components.","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"]