[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-117605-en":3,"doc-seo-117605-105":30,"detail-sidebar-cat-0-en-105":92},{"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":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},117605,962075114765,"Quinn","https://ap-avatar.wpscdn.com/davatar_a8503ba1806abce46bf441b54a3ca4cd",8,"Research & Report","Distribution-Free and Calibrated Predictive Uncertainty in Probabilistic Machine Learning","Machine learning models increasingly support high-stakes decisions in domains such as healthcare and autonomous systems, where wrong predictions can create significant risks. Probabilistic machine learning addresses this by quantifying predictive uncertainty through probabilistic predictions, including predictive distributions and prediction sets with pre-specified coverage. Reliability requires calibration and sharpness: statistical consistency with observed outcomes and concentration around the true target. This dissertation develops distribution-free regression methods using neural networks, including recalibration, regularization, and conformal prediction with finite-sample guarantees.","Université de Mons  \nFaculté des Sciences  \nF.R.S.-FNRS  \nDistribution-Free and Calibrated Predictive Uncertainty in Probabilistic Machine Learning  \nVictor Dheur  \nA dissertation submitted in fulfillment of the requirements of the degree of Docteur en Sciences  \nDecember 2025  \nAdvisor  \nProf. Souhaib Ben Taieb  \nMohamed bin Zayed University of Artificial Intelligence, United Arab Emirates University of Mons, Belgium  \nCo-Advisor  \nProf. Stéphane Dupont  \nUniversity of Mons, Belgium Members of the Jury  \nProf. Jef Wijsen  \nUniversity of Mons, Belgium  \nProf. Pierre Geurts  \nUniversity of Liège, Belgium  \nProf. Matteo Sesia  \nUniversity of Southern California, USA  \nAbstract  \nMachine learning models are increasingly deployed in high-stakes domains such as healthcare and autonomous systems, where decisions carry significant risks. Probabilistic machine learning is valuable in these settings, as it quantifies predictive uncertainty, notably by generating probabilistic predictions. We focus on regression, where the goal is to predict one or more continuous outputs given a set of inputs. In this context, we consider two main forms of uncertainty representation: predictive distributions, which assign probabilities to possible output values, and prediction sets, which are designed to contain the true output with a pre-specified probability. For these predictions to be reliable and informative, they must be calibrated and sharp, i.e., statistically consistent with observed data and concentrated around the true value.  \nIn this thesis, we develop distribution-free regression methods to produce calibrated and sharp probabilistic predictions using neural network models. We consider both single-output and the less-explored multi-output regression settings. Specifically, we develop and study recalibration, regularization, and conformal prediction (CP) methods. The first adjusts predictions after model training, the second augments the training objective, and the last produces prediction sets with finite-sample coverage guarantees.  \nFor single-output regression, we conduct a large-scale experimental study to provide a comprehensive comparison of these methods. The results reveal that post-hoc approaches consistently achieve superior calibration. We explain this finding by establishing a formal link between recalibration and CP, showing that recalibration also benefits from finite-sample coverage guarantees. However, the separate training and recalibration steps typically lead to degraded negative log-likelihood. To address this issue, we develop an end-to-end training procedure that incorporates the recalibration objective directly into learning, resulting in improved negative log-likelihood while maintaining calibration.  \nFor multi-output regression, we conduct a comparative study of CP methods and introduce new classes of approaches that offer novel trade-offs between sharpness, compatibility with generative models, and computational efficiency. A key challenge in CP is achieving conditional coverage, which ensures that coverage guarantees hold for specific inputs rather than only on average. To address this, we propose a method that improves conditional coverage using conditional quantile regression, thereby avoiding the need to estimate full conditional distributions. Finally, for tasks requiring a full predictive density, we introduce a recalibration technique that operates in the latent space of invertible generative models such as conditional normalizing flows. This approach yieldsan explicit, calibrated multivariate probability density function. Collectively, these contributions advance the theory and practice of uncertainty quantification in machine learning, facilitating the development of more reliable predictive systems across diverse applications.  \niv  \nAcknowledgment  \nThis thesis concludes a four-year endeavor, during which I have had the pleasure of working with many remarkable people.  \nI would first like to express my","cbCailhnX5M04rD2","https://ap.wps.com/l/cbCailhnX5M04rD2","pdf",25179089,1,304,"English","en",105,"# Introduction\n## Uncertainty quantification in probabilistic prediction\n# Probabilistic regression and uncertainty representations\n## Predictive distributions\n## Prediction sets\n# Distribution-free calibrated methods\n## Recalibration\n## Regularization\n## Conformal prediction\n# Single-output regression\n## Post-hoc calibration approaches\n## Link between recalibration and conformal prediction\n## End-to-end recalibration training\n# Multi-output regression\n## Conformal prediction for multi-output settings\n## Conditional coverage\n## Conditional quantile regression\n## Latent-space recalibration with generative models","[{\"question\":\"What types of predictive uncertainty does the thesis study in probabilistic machine learning regression?\",\"answer\":\"It studies predictive distributions, which assign probabilities to output values, and prediction sets, designed to contain the true output with a pre-specified probability.\"},{\"question\":\"How does the thesis aim to ensure probabilistic predictions are reliable?\",\"answer\":\"It targets calibration and sharpness, meaning statistical consistency with observed data and concentration of predictions around the true value.\"},{\"question\":\"What are the main contributions for single-output regression?\",\"answer\":\"It develops distribution-free methods using neural networks, shows post-hoc recalibration improves calibration, relates recalibration to finite-sample coverage, and proposes an end-to-end training procedure to improve negative log-likelihood while maintaining calibration.\"}]","Distribution-Free and Calibrated Predictive Uncertainty in Probabilistic Machine Learning | PDF",1785677246,766,{"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":87,"head_meta":89,"extra_data":91,"updated_unix":28},"distribution-free-and-calibrated-predictive-uncertainty-in-probabilistic-machine-learning","",{"@graph":36,"@context":86},[37,54,69],{"@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/distribution-free-and-calibrated-predictive-uncertainty-in-probabilistic-machine-learning/117605/",4,{"url":52,"name":13,"@type":55,"author":56,"headline":13,"publisher":58,"fileFormat":61,"inLanguage":23,"description":14,"dateModified":62,"datePublished":63,"encodingFormat":61,"isAccessibleForFree":64,"interactionStatistic":65},"DigitalDocument",{"name":9,"@type":57},"Person",{"url":41,"name":59,"@type":60},"DocShare","Organization","application/pdf","2026-08-05","2026-08-02",true,{"@type":66,"interactionType":67,"userInteractionCount":20},"InteractionCounter",{"@type":68},"ViewAction",{"@type":70,"mainEntity":71},"FAQPage",[72,78,82],{"name":73,"@type":74,"acceptedAnswer":75},"What types of predictive uncertainty does the thesis study in probabilistic machine learning regression?","Question",{"text":76,"@type":77},"It studies predictive distributions, which assign probabilities to output values, and prediction sets, designed to contain the true output with a pre-specified probability.","Answer",{"name":79,"@type":74,"acceptedAnswer":80},"How does the thesis aim to ensure probabilistic predictions are reliable?",{"text":81,"@type":77},"It targets calibration and sharpness, meaning statistical consistency with observed data and concentration of predictions around the true value.",{"name":83,"@type":74,"acceptedAnswer":84},"What are the main contributions for single-output regression?",{"text":85,"@type":77},"It develops distribution-free methods using neural networks, shows post-hoc recalibration improves calibration, relates recalibration to finite-sample coverage, and proposes an end-to-end training procedure to improve negative log-likelihood while maintaining calibration.","https://schema.org",{"og:url":52,"og:type":88,"og:title":13,"og:site_name":59,"og:description":14},"article",{"robots":90,"canonical":52},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":93},[94,98,102,106,111,116,121,124,129,132,136],{"id":20,"doc_module":4,"doc_module_name":46,"category_name":95,"show_sort_weight":96,"slug":97},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":99,"show_sort_weight":100,"slug":101},"Literature",80,"literature",{"id":53,"doc_module":4,"doc_module_name":46,"category_name":103,"show_sort_weight":104,"slug":105},"Exam",70,"exam",{"id":107,"doc_module":4,"doc_module_name":46,"category_name":108,"show_sort_weight":109,"slug":110},5,"Comic",60,"comic",{"id":112,"doc_module":4,"doc_module_name":46,"category_name":113,"show_sort_weight":114,"slug":115},6,"Technology",50,"technology",{"id":117,"doc_module":4,"doc_module_name":46,"category_name":118,"show_sort_weight":119,"slug":120},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":122,"slug":123},30,"research-report",{"id":125,"doc_module":4,"doc_module_name":46,"category_name":126,"show_sort_weight":127,"slug":128},9,"Religion & Spirituality",20,"religion-spirituality",{"id":127,"doc_module":4,"doc_module_name":46,"category_name":130,"show_sort_weight":127,"slug":131},"World Cup","world-cup",{"id":133,"doc_module":4,"doc_module_name":46,"category_name":134,"show_sort_weight":133,"slug":135},10,"Lifestyle","lifestyle",{"id":137,"doc_module":4,"doc_module_name":46,"category_name":138,"show_sort_weight":107,"slug":139},19,"General","general"]