[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-118931-en":3,"doc-seo-118931-105":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},118931,3848291630094,"Emma Wilson","https://eur-avatar.wpscdn.com/davatar_085a072bc5b1113ac321206ff7593b45",8,"Research & Report","Shift happens - How can machine learning systems be best prepared? - Dissertation","Machine learning systems have achieved major public milestones, yet they remain unreliable under real-world distribution shifts where test data differs from training conditions. This fragility arises from a mismatch between textbook assumptions and deployment reality: models are often developed in one setting and deployed in another, producing failures that are especially dangerous in safety-critical contexts such as healthcare and autonomous driving. The thesis argues that shift is inevitable and presents four approaches to better prepare and adapt systems to such changes, improving reliability for deployment.","This thesis has been submitted in fulfilment of the requirements for a postgraduate degree (e. g. PhD, MPhil, DClinPsychol) at the University of Edinburgh. Please note the following terms and conditions of use:  \n• This work is protected by copyright and other intellectual property rights, which are retained by the thesis author, unless otherwise stated.  \n• A copy can be downloaded for personal non-commercial research or study, without prior permission or charge.  \n• This thesis cannot be reproduced or quoted extensively from without first obtaining permission in writing from the author.  \n• The content must not be changed in any way or sold commercially in any format or medium without the formal permission of the author.  \n• When referring to this work, full bibliographic details including the author, title, awarding institution and date of the thesis must be given.  \nShift happens: How can machine learning systems be best prepared?  \nCian Eastwood  \nU  \nR  \nG  \nH  \nO  \nF  \nE  \nD  \nDoctor of Philosophy  \nInstitute for Adaptive and Neural Computation School of Informatics  \nThe University of Edinburgh  \nAbstract  \nMachine learning systems have made headlines in recent years, defeating world champions in Go, enhancing medical diagnoses, and redefining how we work with tools like ChatGPT. However, despite these impressive feats, machine learning systems remain fragile when faced with test data that differs from their training data. This fragility stems from a fundamental mismatch between textbook machine-learning methods and their real-world application. While textbook methods assume that the conditions under which a system is developed are similar to those in which it is deployed, in reality, systems tend to be developed under one set of conditions (e.g., in a lab) and deployed to another (e.g., a clinic) . As a result, many machine learning systems are not prepared for the condition differences or distribution shifts they face upon deployment, leading to some high-profile and costly failures. For safety-critical settings like healthcare and autonomous driving, such failures represent a major barrier to real-world deployment.  \nIn this thesis, I argue that we must first accept that shift happens, and subsequently focus on how we can best prepare. To do so, I present four of my works that illustrate how machine learning systems can be prepared for (and adapted to) real-world distribution shifts. Together, these contributions take us closer to reliable machine learning systems that can be deployed in safety-critical settings.  \nIn the first work, the setting is source-free domain adaptation, i.e., adapting a model to unlabelled test data without the original training data. Here, we prepare for a change in measurement device (e.g., X-rays from a different scanner) by storing lightweight statistics of the training data. By restoring these statistics on the test data, we see improved accuracy, calibration and data efficiency over prior methods.  \nIn the second work, the setting is domain generalisation, i.e., performing well on test data from new environments or domains by leveraging data from multiple related domains at training time. Here, we prepare for more flexible and unknown changes by exploiting invariances across the training domains that hold with high probability in unseen test domains. In particular, by minimising a particular quantile of a model’s performance distribution over domains, we learn models that perform well with the corresponding probability.  \nIn the third work, the setting is again domain generalisation, but this time we focus on ways to harness so-called “spurious” features without test-domain labels. In particular, we show that predictions based on invariant/stable features can be used to adapt our usage of spurious/unstable features to new test domains, so long as the stable and  \nunstable features are complementary (i.e., conditionally independent given the label) . By safely harnessing complementary ","cbCaimywzRNyJyao","https://ap.wps.com/l/cbCaimywzRNyJyao","pdf",14119354,1,193,"English","en",105,"# Abstract\n## Source-free domain adaptation\n## Domain generalisation via invariances\n## Harnessing complementary spurious features\n## Disentangled representation learning and identifiability\n## Lay summary","[{\"question\":\"Why do machine learning systems fail under real-world distribution shifts?\",\"answer\":\"They are fragile when test data conditions differ from training data. The thesis attributes this to a mismatch between textbook development assumptions and real deployment settings, where conditions often change.\"},{\"question\":\"What does source-free domain adaptation aim to accomplish in the thesis?\",\"answer\":\"It adapts a model to unlabeled test data without access to the original training data. The approach prepares for changes in measurement devices by storing lightweight training statistics and restoring them on test data.\"},{\"question\":\"How does the thesis improve robustness when spurious features are present?\",\"answer\":\"It focuses on using predictions from invariant or stable features to adapt reliance on spurious or unstable features in new test domains. Robust performance is boosted while maintaining safety through a complementary (conditionally independent given the label) relationship between stable and unstable features.\"}]","Shift happens - How can machine learning systems be best prepared? - Dissertation | PDF",1785721026,486,{"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},"shift-happens-how-can-machine-learning-systems-be-best-prepared-dissertation","",{"@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/shift-happens-how-can-machine-learning-systems-be-best-prepared-dissertation/118931/",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},"Why do machine learning systems fail under real-world distribution shifts?","Question",{"text":75,"@type":76},"They are fragile when test data conditions differ from training data. The thesis attributes this to a mismatch between textbook development assumptions and real deployment settings, where conditions often change.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"What does source-free domain adaptation aim to accomplish in the thesis?",{"text":80,"@type":76},"It adapts a model to unlabeled test data without access to the original training data. The approach prepares for changes in measurement devices by storing lightweight training statistics and restoring them on test data.",{"name":82,"@type":73,"acceptedAnswer":83},"How does the thesis improve robustness when spurious features are present?",{"text":84,"@type":76},"It focuses on using predictions from invariant or stable features to adapt reliance on spurious or unstable features in new test domains. Robust performance is boosted while maintaining safety through a complementary (conditionally independent given the label) relationship between stable and unstable features.","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"]