[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-116828-en":3,"doc-seo-116828-105":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},116828,1099513958607,"Jiven","https://ap-avatar.wpscdn.com/avatar/100002390cf8733938c?x-image-process=image/resize,m_fixed,w_180,h_180&k=1778829742770036399",8,"Research & Report","Machine Learning with a Reject Option - A survey","Machine learning models always output a prediction, even when the result is likely to be inaccurate, which is harmful in decision-support settings where errors can carry severe consequences. Machine learning with a reject option addresses this by allowing models to abstain when they expect a mistake. This survey provides an overview of the field, introducing rejection conditions via ambiguity and novelty, specifying model architectures, outlining standard training strategies, connecting traditional learning techniques to rejection, and presenting evaluation methods for predictive and rejective quality, plus application-domain examples and links to related research areas.","View metadata, citation and similar [papers at ](papers at core.ac.uk)[core.ac.uk](papers at core.ac.uk) brought to you by CORE  \n[provided by](provided by arXiv.org)[ arXiv.org](provided by arXiv.org) e-Print Archive  \narXiv :2 107 . 1 1277v 1 [ cs .LG] 23 Jul 2021  \nMachine Learning with a Reject Option: A survey  \nHendrickx Kilian 􀀁 Perini Lorenzo 􀀁 Van der  \nPlas Dries 􀀁 Meert Wannes 􀀁 Davis Jesse  \nReceived: 23 July 2021  \nAbstract Machine learning models always make a prediction, even when it is likely to be inaccurate. This behavior should be avoided in many decision support applications, where mistakes can have severe consequences. Albeit already studied in 1970, machine learning with a reject option recently gained interest. This machine learning sub􀀌eld enables machine learning models to abstain from making a prediction when likely to make a mistake.  \nThis survey aims to provide an overview on machine learning with a reject option. We introduce the conditions leading to two types of rejection, ambiguity and novelty rejection. Moreover, we de􀀌ne the existing architectures for models with a reject option, describe the standard learning strategies to train such models and relate traditional machine learning techniques to rejection. Additionally, wereview strategies to evaluate a model's predictive and rejective quality. Finally, we provide examples of relevant application domains and show how machine learning with rejection relates to other machine learning research areas.  \nK. Hendrickx  \nSiemens Digital Industries Software, Leuven, Belgium KU Leuven, Leuven, Belgium  \nE-mail: [kilian.hendrickx@siemens.com](kilian.hendrickx@siemens.com)  \nE-mail: [kilian.hendrickx@cs.kuleuven.be](kilian.hendrickx@cs.kuleuven.be)  \nL. Perini  \nKU Leuven, Leuven, Belgium  \nE-mail: [lorenzo.perini@cs.kuleuven.be](lorenzo.perini@cs.kuleuven.be)  \nD. Van der Plas  \nOSG bv, Micromed Group, Kontich, Belgium KU Leuven, Leuven, Belgium  \nUniversity of Antwerp, Antwerp, Belgium E-mail: [dries.vanderplas@cs.kuleuven.be](dries.vanderplas@cs.kuleuven.be)  \nW. Meert  \nKU Leuven; Leuven.AI, Leuven, Belgium [E-mail: wannes.meert@cs.kuleuven.be](E-mail: wannes.meert@cs.kuleuven.be)  \nJ. Davis  \nKU Leuven; Leuven.AI, Leuven, Belgium [E-mail: jesse.davis@cs.kuleuven.be](E-mail: jesse.davis@cs.kuleuven.be)  \nKeywords Machine learning with rejection 􀀁 Supervised learning 􀀁 Trustworthy machine learning  \nMathematics Subject Classi􀀌cation (2020) 68T05 􀀁 68T02  \nContents  \n1 Introduction ......................................... 3  \n2 Preliminaries on machine learning with a reject option ................. 4  \n3 Rejection types ....................................... 5  \n4 Properties of a rejection model .............................. 7  \n5 Evaluating models with a reject option .......................... 14  \n6 Instantiating a model with a reject architecture ..................... 16  \n7 Learning a model with a reject option .......................... 23  \n8 Relation between model types and rejection ....................... 28  \n9 Applications of machine learning models that reject .................. 31  \n10 Link to other research areas ................................ 32  \n11 Conclusions and perspectives ............................... 35  \n1 Introduction  \nThe canonical task in machine learning is to learn a predictive model that captures the relationship between a set of input variables and a target variable on the basis of training data. Machine learned models are powerful because after training, they o􀀋er the ability to make accurate predictions about future examples. Since this enables automating a number of tasks that are di􀀎cult and/or time-consuming, such models are ubiquitously deployed.  \nHowever, their key functionality of always returning a prediction for a given novel input is also a drawback. While the model may produce accurate predictions in general, in certain circumstances this may not be the case. For example, there could be certain regi","cbCaitokTbFfi6V2","https://ap.wps.com/l/cbCaitokTbFfi6V2","pdf",719840,1,45,"English","en",105,"# Introduction\n# Preliminaries on machine learning with a reject option\n# Rejection types\n# Properties of a rejection model\n# Evaluating models with a reject option\n# Instantiating a model with a reject architecture\n# Learning a model with a reject option\n# Relation between model types and rejection\n# Applications of machine learning models that reject\n# Link to other research areas\n# Conclusions and perspectives","[{\"question\":\"Why do machine learning models with a reject option matter in decision-support applications?\",\"answer\":\"They prevent unreliable predictions by allowing the model to abstain when it is likely to be wrong. This avoids serious errors in domains where mistakes have severe consequences.\"},{\"question\":\"What are the two types of rejection discussed in the survey?\",\"answer\":\"The survey introduces rejection conditions leading to two types: ambiguity rejection and novelty rejection.\"},{\"question\":\"How are models with a reject option evaluated?\",\"answer\":\"Evaluation focuses on both predictive quality and rejective quality, using strategies designed to measure how well the model performs when it predicts and how effectively it abstains when uncertain.\"}]","Machine Learning with a Reject Option - A survey | PDF",1785671971,113,{"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},"machine-learning-with-a-reject-option-a-survey","",{"@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/machine-learning-with-a-reject-option-a-survey/116828/",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},"Why do machine learning models with a reject option matter in decision-support applications?","Question",{"text":75,"@type":76},"They prevent unreliable predictions by allowing the model to abstain when it is likely to be wrong. This avoids serious errors in domains where mistakes have severe consequences.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"What are the two types of rejection discussed in the survey?",{"text":80,"@type":76},"The survey introduces rejection conditions leading to two types: ambiguity rejection and novelty rejection.",{"name":82,"@type":73,"acceptedAnswer":83},"How are models with a reject option evaluated?",{"text":84,"@type":76},"Evaluation focuses on both predictive quality and rejective quality, using strategies designed to measure how well the model performs when it predicts and how effectively it abstains when uncertain.","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"]