[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-123391-en":3,"doc-seo-123391-105":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},123391,962075114101,"Seraphina","https://ap-avatar.wpscdn.com/avatar/e000253a75eb197efd?x-image-process=image/resize,m_fixed,w_180,h_180&k=1780044092746381165",8,"Research & Report","Privacy meets Robustness - Unveiling the interplay between Differential Privacy and Robustness in Machine Learning","This dissertation studies how differential privacy (DP) can be reconciled with outlier robustness in machine learning under data from large-scale datasets. It examines canonical statistical problems to reveal shared structure between privacy protection and robust estimation, focusing on algorithms that achieve both goals without extra data. The work first introduces an efficient approach with sub-optimal sample complexity, then proposes a unifying framework with nearly optimal sample complexity across mean estimation, linear regression, covariance estimation, and PCA, and finally presents efficient near-optimal methods for differentially private PCA and label-robust linear regression.","©Copyright 2024 Xiyang Liu  \nPrivacy meets Robustness: Unveiling the interplay between Differential Privacy and Robustness in Machine Learning  \nXiyang Liu  \nA dissertation  \nsubmitted in partial fulfillment of the requirements for the degree of  \nDoctor of Philosophy  \nUniversity of Washington  \n2024  \nReading Committee:  \nSewoong Oh, Chair  \nSimon Du  \nPang Wei Koh  \nProgram Authorized to Offer Degree:  \nComputer Science and Engineering  \nUniversity of Washington  \nAbstract  \nPrivacy meets Robustness: Unveiling the interplay between Differential Privacy and  \nRobustness in Machine Learning  \nXiyang Liu  \nChair of the Supervisory Committee:  \nProfessor Sewoong Oh  \nPaul G. Allen School of Computer Science and Engineering  \nThe rapid advancement of machine learning over the past decade has been driven by the increasing availability of large-scale datasets. However, this growth has raised critical concerns regarding the privacy of individuals whose data is being used, as well as the robustness of algorithms against potentially malicious data corruption from unreliable sources. This thesis aims to explore the fundamental interplay between differential privacy (DP) and outlier robustness in machine learning.  \nThis thesis investigates several canonical statistical problems to uncover the inherent connections between DP and robustness. The first problem addresses whether it is possible to develop algorithms that are both differentially private and robust to outliers without requiring additional data. We present the first efficient algorithm with sub-optimal sample complexity. Then, we introduce a unifying framework that achieves nearly optimal sample complexity, without considering computational efficiency, across various problems, including mean estimation, linear regression, covariance estimation, and principal component analysis (PCA) . Finally, we propose two efficient algorithms that achieve near-optimal sample complexity for differentially private PCA and linear regression.  \nThe findings of this research contribute to a deeper understanding of the interplay between privacy and robustness, providing new insights into the design of algorithms that are both  \nstatistically optimal and computationally efficient for practical applications. The results presented in this thesis open avenues for further exploration into the protection of data privacy, particularly in high-dimensional and adversarial settings.  \nTABLE OF CONTENTS  \nPage  \nList of Figures ....................................... iv  \nChapter 1: Introduction ................................ 1  \n1.1 Preliminaries ................................... 3  \nChapter 2: Differentially private and robust mean estimation ............ 5  \n2.1 Introduction .................................... 5  \n2.2 Background on exponential time approaches for Gaussian distributions ... 14  \n2.3 Efficient algorithms for private and robust mean estimation .......... 16  \n2.4 Exponential time approaches for sub-Gaussian distributions ......... 31  \n2.5 Heavy-tailed distributions: algorithm and analysis ............... 33  \n2.6 Discussion ..................................... 34  \nChapter 3: HPTR: A unifying framework for differentially private and robust estimation 36  \n3.1 Introduction .................................... 36  \n3.2 Preliminaries ................................... 55  \n3.3 Mean estimation ................................. 59  \n3.4 Linear regression ................................. 86  \n3.5 Covariance estimation .............................. 110  \n3.6 Principal component analysis .......................... 117  \n3.7 Discussion ..................................... 130  \nChapter 4: Differentially private PCA ......................... 132  \n4.1 Introduction .................................... 132  \n4.2 Problem formulation and background on DP .................. 134  \n4.3 First attempt: making Oja’s Algorithm private ................. 137  \n4.4 Two remaining challenges ............","cbCaioeLleGzPIz4","https://ap.wps.com/l/cbCaioeLleGzPIz4","pdf",2305494,1,327,"English","en",105,"# Abstract\n# Introduction\n## Preliminaries\n# Differentially private and robust mean estimation\n## Background and efficient algorithms\n## Heavy-tailed distributions and discussion\n# HPTR: A unifying framework for differentially private and robust estimation\n## Mean estimation, linear regression, covariance estimation, PCA\n## Discussion\n# Differentially private PCA\n## DP formulation, Oja’s algorithm privacy, and challenges\n# Label-robust differentially private linear regression\n## Problem setup, formulation, and experimental results\n# Bibliography\n# Appendices","[{\"question\":\"What central problem does this dissertation address?\",\"answer\":\"It investigates the interplay between differential privacy and outlier robustness in machine learning, aiming to understand and design algorithms that satisfy both simultaneously.\"},{\"question\":\"How does the thesis approach canonical statistical problems?\",\"answer\":\"It studies multiple problems—mean estimation, linear regression, covariance estimation, and PCA—using frameworks and algorithms that connect DP requirements with robustness to outliers.\"},{\"question\":\"What algorithmic contributions are presented for PCA and linear regression?\",\"answer\":\"The thesis proposes two efficient algorithms with near-optimal sample complexity for differentially private PCA and for label-robust differentially private linear regression.\"}]","Privacy meets Robustness - Unveiling the interplay between Differential Privacy and Robustness in Machine Learning | PDF",1785816250,824,{"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},"privacy-meets-robustness-unveiling-the-interplay-between-differential-privacy-and-robustness-in-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/privacy-meets-robustness-unveiling-the-interplay-between-differential-privacy-and-robustness-in-machine-learning/123391/",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 central problem does this dissertation address?","Question",{"text":75,"@type":76},"It investigates the interplay between differential privacy and outlier robustness in machine learning, aiming to understand and design algorithms that satisfy both simultaneously.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How does the thesis approach canonical statistical problems?",{"text":80,"@type":76},"It studies multiple problems—mean estimation, linear regression, covariance estimation, and PCA—using frameworks and algorithms that connect DP requirements with robustness to outliers.",{"name":82,"@type":73,"acceptedAnswer":83},"What algorithmic contributions are presented for PCA and linear regression?",{"text":84,"@type":76},"The thesis proposes two efficient algorithms with near-optimal sample complexity for differentially private PCA and for label-robust differentially private linear regression.","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"]