[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-128429-en":3,"doc-seo-128429-105":31,"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":21,"is_downloadable":21,"audit_status":21,"page_count":22,"language":23,"language_code":24,"site_id":25,"html_lang":24,"table_of_contents":26,"faqs":27,"seo_title":28,"seo_description":14,"update_tm":29,"read_time":30},128429,962084925290,"Ophelia","https://ap-avatar.wpscdn.com/davatar_085a072bc5b1113ac321206ff7593b45",8,"Research & Report","On Learning, Fairness, and Complexity - Thesis","Study of learning and complexity-theoretic foundations of the multigroup fairness framework for prediction algorithms. The work analyzes multigroup notions of multiaccuracy and multicalibration, showing they can be achieved via weak agnostic learning and that multicalibration implies omniprediction. The thesis connects these fairness notions to complexity theory through the Regularity Lemma and the Hardcore Lemma, proving how calibrated multiaccuracy recovers stronger agnostic learning and optimal hardcore density. It also develops selective omnipredictors and abstention methods with fairness and conformal coverage guarantees.","On Learning, Fairness, and Complexity  \nSílvia Casacuberta Puig  \nSt John’s College University of Oxford  \nA thesis submitted for the degree of Master of Science (by Research) in Computer Science  \nTrinity 2025  \nii  \nThesis advisor: Prof. Varun Kanade Sílvia Casacuberta Puig  \nOn Learning, Fairness, and Complexity  \nAbstract  \nIn this thesis we study the learning and complexity-theoretic underpinnings of the multigroup fairness framework for prediction algorithms. Multiaccuracy and multicalibration are two primary multigroup fairness notions, which ensure accurate and calibrated predictions, respectively, for every subpopulation that can be identified within a specified class of computations [HKRR18] . They both can be achieved from a single learning primitive: weak agnostic learning. A line of work starting from [GKR+22] has shown that multicalibration implies a very strong indistinguishability-based form of learning called omniprediction. The multigroup fairness framework is also deeply connected to complexity theory through the Regularity Lemma and its various implications [CDV24] .  \nWe provide a thorough study of the connections between multigroup fairness notions, the central learning primitive of weak agnostic learning, and the fundamental Hardcore Lemma in complexity theory. We find that multiaccuracy in itself is rather weak, but that the addition of global calibration (this notion is called calibrated multiaccuracy) boosts its power substantially, enough to recover implications that were previously known only assuming the stronger notion of multicalibration.  \nWe give evidence that multiaccuracy might not be as powerful as standard weak agnostic learning, by showing that there is no way to post-process a multiaccurate predictor to get a weak learner, even assuming the best hypothesis has correlation 1/2 . However, by also requiring the predictor to be calibrated, we recover not just weak, but strong agnostic learning. A similar picture emerges when we consider the derivation of hardcore measures from predictors satisfying multigroup fairness notions [TTV09; CDV24] . On the one hand, while multiaccuracy only yields hardcore measures of density half the optimal, we show that (a weighted version of) calibrated multiaccuracy achieves optimal density.  \nOur results yield new insights into the complementary roles played by multiaccuracy and calibration in each setting. They shed light on why multiaccuracy and global calibration, although not particularly powerful by themselves, together yield considerably stronger notions.  \nWe further study the connections between the multigroup fairness framework and the problem of learning selective classifiers, which are predictors that are allowed to abstain on some fraction of the domain. Building on the notion of omniprediction (which is in turn built using tools from the multigroup fairness framework), given a pre-specified class of loss functions, we provide an algorithm for efficiently building a single classifier that learns abstentions and predictions optimally for every loss in the entire class, where the abstentions are decided efficiently for each specific loss function by applying a fixed post-processing function. We call this classifier a selective omnipredictor. Our algorithm and theoretical guarantees generalize the previously-known algorithms for learning selective classifiers in formal learning-theoretic models [KKM12] .  \nThesis advisor: Prof. Varun Kanade Sílvia Casacuberta Puig  \nWe then extend the traditional multigroup fairness algorithms to the selective classification setting and show that we can use a calibrated and multiaccurate predictor to efficiently build selective classifiers that abstain optimally not only globally but also locally within each of the groups in any pre-specified collection of possibly intersecting subgroups of the domain, and are also accurate when they do not abstain. This provides yet another use case of the notion of calibrated multiaccuracy. Moreover, ","cbCaisI0cN3k7RuD","https://ap.wps.com/l/cbCaisI0cN3k7RuD","pdf",4519005,2,1,178,"English","en",105,"# Contents\n## 1 Introduction\n## 2 Notation & Preliminaries\n## I Agnostic Learning, Multigroup Fairness, and Hardcore Measures\n## 3 Multiaccuracy & Agnostic Learning\n## 4 Impagliazzo’s Hardcore Lemma","[{\"question\":\"What fairness and learning notions are central to the thesis?\",\"answer\":\"The thesis focuses on multigroup fairness notions including multiaccuracy and multicalibration, and their relation to learning through weak agnostic learning.\"},{\"question\":\"How does multicalibration relate to a stronger learning concept?\",\"answer\":\"Starting from prior results, the thesis explains that multicalibration implies omniprediction, a strong indistinguishability-based learning form.\"},{\"question\":\"Why does calibrated multiaccuracy strengthen the results compared with multiaccuracy alone?\",\"answer\":\"Multiaccuracy by itself is shown to be relatively weak, but adding global calibration yields calibrated multiaccuracy with substantially stronger implications, recovering results previously known only from multicalibration.\"}]","On Learning, Fairness, and Complexity - Thesis | PDF",1785947632,449,{"code":4,"msg":32,"data":33},"ok",{"site_id":25,"language":24,"slug":34,"title":13,"keywords":35,"description":14,"schema_data":36,"social_meta":87,"head_meta":89,"extra_data":91,"updated_unix":29},"on-learning-fairness-and-complexity-thesis","",{"@graph":37,"@context":86},[38,54,69],{"@type":39,"itemListElement":40},"BreadcrumbList",[41,45,48,51],{"item":42,"name":43,"@type":44,"position":21},"https://docshare.wps.com","Home","ListItem",{"item":46,"name":47,"@type":44,"position":20},"https://docshare.wps.com/document/","Document",{"item":49,"name":12,"@type":44,"position":50},"https://docshare.wps.com/document/research-report/",3,{"item":52,"name":13,"@type":44,"position":53},"https://docshare.wps.com/document/on-learning-fairness-and-complexity-thesis/128429/",4,{"url":52,"name":13,"@type":55,"author":56,"headline":13,"publisher":58,"fileFormat":61,"inLanguage":24,"description":14,"dateModified":62,"datePublished":63,"encodingFormat":61,"isAccessibleForFree":64,"interactionStatistic":65},"DigitalDocument",{"name":9,"@type":57},"Person",{"url":42,"name":59,"@type":60},"DocShare","Organization","application/pdf","2026-08-23","2026-08-05",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 fairness and learning notions are central to the thesis?","Question",{"text":76,"@type":77},"The thesis focuses on multigroup fairness notions including multiaccuracy and multicalibration, and their relation to learning through weak agnostic learning.","Answer",{"name":79,"@type":74,"acceptedAnswer":80},"How does multicalibration relate to a stronger learning concept?",{"text":81,"@type":77},"Starting from prior results, the thesis explains that multicalibration implies omniprediction, a strong indistinguishability-based learning form.",{"name":83,"@type":74,"acceptedAnswer":84},"Why does calibrated multiaccuracy strengthen the results compared with multiaccuracy alone?",{"text":85,"@type":77},"Multiaccuracy by itself is shown to be relatively weak, but adding global calibration yields calibrated multiaccuracy with substantially stronger implications, recovering results previously known only from multicalibration.","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":25},{"code":4,"msg":5,"data":93},[94,98,102,106,111,116,121,124,129,132,136],{"id":21,"doc_module":4,"doc_module_name":47,"category_name":95,"show_sort_weight":96,"slug":97},"Story & Novel",90,"story-novel",{"id":20,"doc_module":4,"doc_module_name":47,"category_name":99,"show_sort_weight":100,"slug":101},"Literature",80,"literature",{"id":53,"doc_module":4,"doc_module_name":47,"category_name":103,"show_sort_weight":104,"slug":105},"Exam",70,"exam",{"id":107,"doc_module":4,"doc_module_name":47,"category_name":108,"show_sort_weight":109,"slug":110},5,"Comic",60,"comic",{"id":112,"doc_module":4,"doc_module_name":47,"category_name":113,"show_sort_weight":114,"slug":115},6,"Technology",50,"technology",{"id":117,"doc_module":4,"doc_module_name":47,"category_name":118,"show_sort_weight":119,"slug":120},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":47,"category_name":12,"show_sort_weight":122,"slug":123},30,"research-report",{"id":125,"doc_module":4,"doc_module_name":47,"category_name":126,"show_sort_weight":127,"slug":128},9,"Religion & Spirituality",20,"religion-spirituality",{"id":127,"doc_module":4,"doc_module_name":47,"category_name":130,"show_sort_weight":127,"slug":131},"World Cup","world-cup",{"id":133,"doc_module":4,"doc_module_name":47,"category_name":134,"show_sort_weight":133,"slug":135},10,"Lifestyle","lifestyle",{"id":137,"doc_module":4,"doc_module_name":47,"category_name":138,"show_sort_weight":107,"slug":139},19,"General","general"]