[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-124841-en":3,"doc-seo-124841-105":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},124841,4398048949847,"Eliana","https://ap-avatar.wpscdn.com/avatar/400002536579ef2da7f?_k=1778318612642679267",8,"Research & Report","Why and How to Construct an Epistemic Justification of Machine Learning","The paper examines why inductive generalizations are learnable from shuffled observations drawn from a fixed probability distribution, even though this learning is inverse and ill-posed. It argues that complexity regularization can provide an epistemic justification when it satisfies Norton’s Material Theory of Induction by localizing inductive risk to facts within the given domain. It uses the Lottery Ticket Hypothesis, stability analysis of Empirical Risk Minimization, and local stability to avoid failures under asymptotic instability. The work clarifies a division of labor between MTI and optimality-based justifications for neural networks.","Why and How to Construct an Epistemic Justification of Machine Learning?  \nPetr Spelda (first and corresponding author, [petr.spelda@fsv.cuni.cz](petr.spelda@fsv.cuni.cz) ), Vit Stritecky (second author); affiliation: Department of Security Studies, Institute of Political Studies, Faculty of Social Sciences, Charles University.  \nAccepted for publication in Synthese (DOI: 10. 1007/s11229-024-04702-z) .  \nAbstract  \nConsider a set of shuffled observations drawn from a fixed probability distribution over some instance domain. What enables learning of inductive generalizations which proceed from such a set of observations? The scenario is worthwhile because it epistemically characterizes most of machine learning. This kind of learning from observations is also inverse and ill-posed. What reduces the non-uniqueness of its result and, thus, its problematic epistemic justification, which stems from a one-to-many relation between the observations and many learnable generalizations? The paper argues that this role belongs to any complexity regularization which satisfies Norton’s Material Theory of Induction (MTI) by localizing the inductive risk to facts in the given domain. A prime example of the localization is the Lottery Ticket Hypothesis (LTH) about overparameterized neural networks. The explanation of MTI’s role in complexity regularization of neural networks is provided by analyzing the stability of Empirical Risk Minimization (ERM), an inductive rule that controls the learning process and leads to an inductive generalization on the given set of observations. In cases where ERM might become asymptotically unstable, making the justification of the generalization by uniform convergence unavailable, LTH and MTI can be used to define a local stability. A priori, overparameterized neural networks are such cases and the combination of LTH and MTI can block ERM’s trivialization caused by equalizing the strengths of its inductive support for risk minimization. We bring closer the investigation of generalization in artificial neural networks and the study of inductive inference and show the  \ndivision of labor between MTI and the optimality justifications (developed by Gerhard Schurz) in machine learning.  \nKeywords: lottery ticket hypothesis; complexity regularization; material theory of induction; empirical risk minimization.  \n1. Introduction  \nAn epistemic justification of the inductive generalization in artificial neural networks can be achieved by connecting the state-of-the-art approaches (LeCun et al. 2015; Schmidhuber 2015; Bengio et al. 2021) to recent theories of inductive inference (Norton 2003; 2021; Schurz 2019) . If machine learning is considered as a kind of induction, then the epistemic justification is missing in machine learning as well as in epistemology debates.  \nIn Sections 2 and 3, the paper connects complexity regularization of (deep) artificial neural networks, the Lottery Ticket Hypothesis (Frankle and Carbin 2019), and the Material Theory of Induction (Norton 2003; 2014; 2021) to show that successful machine learning of inductive generalizations is epistemically justifiable by the localization of inductive risk. Sections 4 and 5 provide important qualifications to this epistemic justification, using Norton’s work on the incompleteness of calculi of inductive inference (2019) to distinguish between asymptotic and local stability of inductive rules that facilitate the generalization learning in neural networks.  \nThe Material Theory of Induction argues that retrodictive/predictive successes of induction stem from adapting general inductive schemas to material facts found in local domains, thus achieving the schemas’ localization. The paper shows that neural network pruning described by the Lottery Ticket Hypothesis adapts a general architecture to a given local domain. By this, it transports the inductive risk from a schema (architecture) to local facts populating the evidence (training data), thus accomplishin","cbCaijqYnOUkZFRK","https://ap.wps.com/l/cbCaijqYnOUkZFRK","pdf",430954,1,42,"English","en",105,"# Introduction\n## Motivation for an Epistemic Justification of Inductive Generalizations\n# Complexity Regularization, LTH, and Material Theory of Induction\n# Qualifications: Asymptotic vs. Local Stability of Inductive Rules\n# Division of Labor and Limitations in Machine Learning Context","[{\"question\":\"What problem does the paper address about machine learning from observations?\",\"answer\":\"It asks what enables learning of inductive generalizations from a set of shuffled observations, despite the inverse and ill-posed nature of this learning.\"},{\"question\":\"How does the Material Theory of Induction relate to complexity regularization?\",\"answer\":\"It supports complexity regularization when inductive risk is localized to facts in the given local domain, which improves the epistemic justification for generalization.\"},{\"question\":\"Why are stability concepts important in the paper’s justification?\",\"answer\":\"Stability of Empirical Risk Minimization determines whether uniform convergence can justify generalization; when ERM is asymptotically unstable, local stability via LTH and MTI is used instead.\"}]","Why and How to Construct an Epistemic Justification of Machine Learning | PDF",1785894944,106,{"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},"why-and-how-to-construct-an-epistemic-justification-of-machine-learning","",{"@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/why-and-how-to-construct-an-epistemic-justification-of-machine-learning/124841/",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","2026-08-05",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 problem does the paper address about machine learning from observations?","Question",{"text":75,"@type":76},"It asks what enables learning of inductive generalizations from a set of shuffled observations, despite the inverse and ill-posed nature of this learning.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How does the Material Theory of Induction relate to complexity regularization?",{"text":80,"@type":76},"It supports complexity regularization when inductive risk is localized to facts in the given local domain, which improves the epistemic justification for generalization.",{"name":82,"@type":73,"acceptedAnswer":83},"Why are stability concepts important in the paper’s justification?",{"text":84,"@type":76},"Stability of Empirical Risk Minimization determines whether uniform convergence can justify generalization; when ERM is asymptotically unstable, local stability via LTH and MTI is used instead.","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"]