[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-122941-en":3,"doc-seo-122941-105":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},122941,4398048950312,"Violet","https://ap-avatar.wpscdn.com/avatar/400002538284de19e3c?_k=1778320343897328908",8,"Research & Report","Rethinking Explainable Machine Learning as Applied Statistics","Rethinking explainable machine learning through the lens of applied statistics reframes what explanation algorithms are for and how they should be used. Explanations are treated as statistics of high-dimensional functions, analogous to traditional statistical quantities, with interpretation as the decisive element. The position paper argues that current research often adds explanation methods without addressing interpretability, leading to recurrent misinterpretation. Drawing on this analogy offers concrete improvements to research practices, evaluation, and benchmarking.","Position: Rethinking Explainable Machine Learning as Applied Statistics  \nSebastian Bordt 1 Eric Raidl 2 Ulrike von Luxburg 1  \narXiv :2402 .02870v5 [ cs .LG] 16 Jun 2025  \nAbstract  \nIn the rapidly growing literature on explanation algorithms, it often remains unclear what precisely these algorithms are for and how they should be used. In this position paper, we argue for a novel and pragmatic perspective: Explainable machine learning needs to recognize its parallels with applied statistics. Concretely, explanations are statistics of high-dimensional functions, and we should think about them analogously to traditional statistical quantities. Among others, this implies that we must think carefully about the matter of interpretation, or how the explanations relate to intuitive questions that humans have about the world.  \nThe fact that this is scarcely being discussed in research papers is one of the main drawbacks of the current literature. Moving forward, the analogy between explainable machine learning and applied statistics suggests fruitful ways for how research practices can be improved.  \n1. Introduction  \nDespite a growing literature on explanation algorithms, their use cases, and evaluation, the field of explainable machine learning remains in a pre-paradigmatic state. This is because there is no agreement on the meaning of the basic terminology in the field—what is an explanation? when is a model interpretable?—as has been repeatedly noted indifferent papers (Doshi-Velez & Kim, 2017 ; Lipton, 2018 ; Murdoch et al., 2019 ; Freiesleben & Knig, 2023) .  \nIn this position paper, we argue for a novel perspective on explainable machine learning: The field needs to recognize its parallels with applied statistics. By applied statistics, we mean the use of statistical theory, methods, and tools to analyze data and solve real-world problems (Wilcox & Rousselet, 2023 ; Angrist & Pischke, 2009 ; Cox & Donnelly,  \n1University of T¨ubingen, T¨ubingen AI Center, Germany 2University of T¨ubingen, Germany. Correspondence to: Sebastian Bordt \u003C[sebastian.bordt@uni-tuebingen.de](sebastian.bordt@uni-tuebingen.de) >.  \nProceedings of the 42 nd International Conference on Machine Learning, Vancouver, Canada. PMLR 267, 2025 . Copyright 2025 by the author(s) .  \n2011 ; Cox, 2018 ; Franconeri et al., 2021) . Both explainable machine learning and applied statistics aim to summarize the behaviour of high-dimensional objects—functions in explainability, and datasets and their respective probability distributions in statistics (Wasserman, 2004) . Moreover, both fields share a common goal: determining how relevant questions about the real world can be answered using such summaries.  \nThe advantage of our perspective is that applied statistics is a fairly established field of research, which means that research practices in explainable machine learning can be improved by drawing appropriate analogies. In this paper, we argue that it is especially important to distinguish between the following two concepts:  \n(a) The mathematical form and properties of an explanation algorithm.  \n(b) The matter of its interpretation.  \nMathematically, explanation algorithms have the form of a functional—a function of functions (Rudin, 1991) . They take a function and map it to a simpler object, often a vector. One might say that the functional “explains” a function by reducing its complexity. However, not all dimensionreducing functionals have explanatory value. For this tobe the case, the functional also needs to formalize an intuitive concept, a nuanced matter debated in the philosophy of science (Justus, 2012) . If this is the case, then we say that the functional has an interpretation. Traditional statistical quantities like the p-value, confidence intervals and visualizations usually have an interpretation.  \nIn the current literature on explainable machine learning, papers often introduce novel explanation algorithms without discussing their interpretation. Thi","cbCaik2mWgaCanhc","https://ap.wps.com/l/cbCaik2mWgaCanhc","pdf",398583,1,13,"English","en",105,"# Abstract\n# Introduction\n## Clarifying terminology in explainable machine learning\n## Applied statistics as the guiding analogy\n## Two key concepts: algorithmic form vs interpretation\n## Risks of explanations without interpretation\n## Recommendations for improving research practices","[{\"question\":\"Why does the paper argue that explainable machine learning is still in a pre-paradigmatic state?\",\"answer\":\"It cites the lack of agreement on key terminology, such as what qualifies as an explanation and when a model is considered interpretable.\"},{\"question\":\"How does the paper define the relationship between explanations and interpretation?\",\"answer\":\"It distinguishes the mathematical form of explanation algorithms from the question of interpretation, arguing that explanatory value requires formalizing an intuitive, human-relevant concept.\"},{\"question\":\"What practical recommendations does the analogy to applied statistics lead to?\",\"answer\":\"The paper recommends designing explanation algorithms to answer specific questions, recognizing the expertise required to use statistics-like outputs, and considering how explanations relate to other measures such as fairness, robustness, and benchmark datasets.\"}]","Rethinking Explainable Machine Learning as Applied Statistics | PDF",1785813796,33,{"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},"rethinking-explainable-machine-learning-as-applied-statistics","",{"@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/rethinking-explainable-machine-learning-as-applied-statistics/122941/",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","2026-08-04",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},"Why does the paper argue that explainable machine learning is still in a pre-paradigmatic state?","Question",{"text":76,"@type":77},"It cites the lack of agreement on key terminology, such as what qualifies as an explanation and when a model is considered interpretable.","Answer",{"name":79,"@type":74,"acceptedAnswer":80},"How does the paper define the relationship between explanations and interpretation?",{"text":81,"@type":77},"It distinguishes the mathematical form of explanation algorithms from the question of interpretation, arguing that explanatory value requires formalizing an intuitive, human-relevant concept.",{"name":83,"@type":74,"acceptedAnswer":84},"What practical recommendations does the analogy to applied statistics lead to?",{"text":85,"@type":77},"The paper recommends designing explanation algorithms to answer specific questions, recognizing the expertise required to use statistics-like outputs, and considering how explanations relate to other measures such as fairness, robustness, and benchmark datasets.","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"]