[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-122978-en":3,"doc-seo-122978-105":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},122978,2336464648322,"Aria","https://ap-avatar.wpscdn.com/avatar/2200025388227c56fec?_k=1778556882303663488",8,"Research & Report","Stability and Hypothesis Transfer Learning - Abstract","Transfer learning is studied in the Hypothesis Transfer Learning (HTL) setting, where the learner lacks direct access to the source domain and instead relies on hypotheses induced from it. A theoretical analysis is conducted for HTL algorithms based on Regularized Least Squares with biased regularization, using algorithmic stability. Results show that when source and target domains are related, Leave-One-Out error converges faster to generalization error, enabling optimal transfer parameter selection even with small training sets.","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  \nprovided by Infoscience- École polytechnique fédérale de Lausanne  \nStability and Hypothesis Transfer Learning  \nIlja Kuzborskij  \nIdiap Research Institute, Switzerland  \n􀀓  \nEcole Polytechnique F􀀓ed􀀓erale de Lausanne (EPFL), Switzerland  \nFrancesco Orabona  \n[ilja.kuzborskij@idiap.ch](ilja.kuzborskij@idiap.ch)  \n[francesco@orabona.com](francesco@orabona.com)  \nToyota Technological Institute at Chicago, USA  \nAbstract  \nWe consider the transfer learning scenario, where the learner does not have access to the source domain directly, but rather operates on the basis of hypotheses induced from it { the Hypothesis Transfer Learning (HTL) problem. Particularly, we conduct a theoretical analysis of HTL by considering the algorithmic stability of a class of HTL algorithms based on Regularized Least Squares with biased regularization. We show that the relatedness of source and target domains accelerates the convergence of the Leave-OneOut error to the generalization error, thus enabling the use of the Leave-One-Out error to  \n􀀌nd the optimal transfer parameters, even in the presence of a small training set. In case of unrelated domains we also suggest a theoretically principled way to prevent negative transfer, so that in the limit we recover the performance of the algorithm not using any knowledge from the source domain.  \n1. Introduction  \nThe standard assumption in supervised machine learning algorithms is to have models trained and tested on samples drawn from the same probability distribution. However, this assumption is often violated in practical applications.  \nA more general setting is the one in which the marginal distributions over training and testing domains are different but related. This is the problem of Domain Adaptation (DA), where a successful scheme typically  \nProceedings of the 30 th International Conference on Machine Learning, Atlanta, Georgia, USA, 2013 . JMLR: W&CP volume 28 . Copyright 2013 by the author(s) .  \nutilizes large unlabeled samples from both domains to adapt a source hypothesis to the target domain. Previous work has addressed in detail the theory of DA and proposed algorithms that critically depend on optimal weighting parameters given by the theoretical analysis (Ben-David et al. , 2010a;b; Mansour et al. , 2009b; Cortes et al. , 2008) . However, in practice, the learner needs access to su􀀎cient unlabeled samples from both domains to estimate these parameters. Even if unlabeled data are abundant, the estimation of these parameters can be computationally prohibitive in some scenarios. A hypothetical example is a large number of domains involved or, for instance, when one acquires new domains incrementally. Here, keeping unlabeled data from all the domains and reestimating parameters is a necessity.  \nTo overcome this practical limitation, a new framework has been analyzed by a number of works (FeiFei et al. , 2006; Yang et al. , 2007; Orabona et al. , 2009; Mansour et al. , 2009a; Tommasi et al. , 2010; Kuzborskij et al. , 2013) . In this framework, that we will call Hypothesis Transfer Learning (HTL), unlike DA, only source hypotheses trained on a source domain are utilized. The attractive quality of HTL is the fact, that it assumes no explicit access to the source domain, nor any knowledge about the relatedness of the source and target distributions. Although, this setting has been explored empirically with success, a formal theory of HTL is mostly missing. Hence it is unclear how to recover optimal transfer parameters and what properties of the source hypothesis a􀀋ect generalization.  \nIn this paper, we take a step towards a theory of HTL. In particular, we analyze the generalization ability of a class of HTL algorithms stemming from Regularized Least Squares (RLS) with biased regularization. We assume access to a given number of source hypotheses and a small set o","cbCaiu3RtbnzBbvk","https://ap.wps.com/l/cbCaiu3RtbnzBbvk","pdf",379900,1,9,"English","en",105,"# Abstract\n# Introduction\n## Domain Adaptation background\n## Hypothesis Transfer Learning framework\n# Definitions and Problem Setting\n# Algorithms and Main Result (Outline)\n## Main theorem and implications\n## Proof approach\n## Related work and conclusions","[{\"question\":\"What is Hypothesis Transfer Learning (HTL) compared with domain adaptation?\",\"answer\":\"HTL uses only source hypotheses trained on a source domain, without requiring direct access to the source data or knowledge of distribution relatedness. Domain adaptation typically adapts using large unlabeled samples from both domains.\"},{\"question\":\"How does the paper analyze HTL theoretically?\",\"answer\":\"It analyzes generalization ability for a class of HTL algorithms derived from Regularized Least Squares with biased regularization, using Leave-One-Out (LOO) risk and hypothesis stability to control differences between expected risk and LOO risk.\"},{\"question\":\"How does relatedness between source and target domains affect learning?\",\"answer\":\"When domains are related, the variance of the LOO estimator decreases as the source hypothesis quality over the target domain increases, accelerating convergence from LOO error to generalization error and supporting reliable transfer parameter selection.\"}]","Stability and Hypothesis Transfer Learning - Abstract | PDF",1785813992,23,{"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},"stability-and-hypothesis-transfer-learning-abstract","",{"@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/stability-and-hypothesis-transfer-learning-abstract/122978/",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","2026-08-04",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 is Hypothesis Transfer Learning (HTL) compared with domain adaptation?","Question",{"text":75,"@type":76},"HTL uses only source hypotheses trained on a source domain, without requiring direct access to the source data or knowledge of distribution relatedness. Domain adaptation typically adapts using large unlabeled samples from both domains.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How does the paper analyze HTL theoretically?",{"text":80,"@type":76},"It analyzes generalization ability for a class of HTL algorithms derived from Regularized Least Squares with biased regularization, using Leave-One-Out (LOO) risk and hypothesis stability to control differences between expected risk and LOO risk.",{"name":82,"@type":73,"acceptedAnswer":83},"How does relatedness between source and target domains affect learning?",{"text":84,"@type":76},"When domains are related, the variance of the LOO estimator decreases as the source hypothesis quality over the target domain increases, accelerating convergence from LOO error to generalization error and supporting reliable transfer parameter selection.","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},[93,97,101,105,110,115,120,123,127,130,134],{"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":21,"doc_module":4,"doc_module_name":46,"category_name":124,"show_sort_weight":125,"slug":126},"Religion & Spirituality",20,"religion-spirituality",{"id":125,"doc_module":4,"doc_module_name":46,"category_name":128,"show_sort_weight":125,"slug":129},"World Cup","world-cup",{"id":131,"doc_module":4,"doc_module_name":46,"category_name":132,"show_sort_weight":131,"slug":133},10,"Lifestyle","lifestyle",{"id":135,"doc_module":4,"doc_module_name":46,"category_name":136,"show_sort_weight":106,"slug":137},19,"General","general"]