Many large-scale machine learning (ML) applications need to perform decentralized learning over datasets generated at different devices and locations. Such datasets pose a significant challenge to decentralized learning because their different contexts result in significant data distribution skew across devices\/locations. In this paper, we take a step toward better understanding this challenge by presenting a detailed experimental study of decentralized DNN training on a common type of data skew: skewed distribution of data labels across devices\/locations. Our study shows that: (i) skewed data labels are a fundamental and pervasive problem for decentralized learning, causing significant accuracy loss across many ML applications, DNN models, training datasets, and decentralized learning algorithms; (ii) the problem is particularly challenging for DNN models with batch normalization; and (iii) the degree of data skew is a key determinant of the difficulty of the problem. Based on these findings, we present SkewScout, a system-level approach that adapts the communication frequency of decentralized learning algorithms to the (skew-induced) accuracy loss between data partitions. We also show that group normalization can recover much of the accuracy loss of batch normalization.<\/p>\n","protected":false},"excerpt":{"rendered":"
Many large-scale machine learning (ML) applications need to perform decentralized learning over datasets generated at different devices and locations. Such datasets pose a significant challenge to decentralized learning because their different contexts result in significant data distribution skew across devices\/locations. In this paper, we take a step toward better understanding this challenge by presenting a […]<\/p>\n","protected":false},"featured_media":0,"template":"","meta":{"msr-url-field":"","msr-podcast-episode":"","msrModifiedDate":"","msrModifiedDateEnabled":false,"ep_exclude_from_search":false,"footnotes":""},"msr-content-type":[3],"msr-research-highlight":[],"research-area":[13556,13547],"msr-publication-type":[193716],"msr-product-type":[],"msr-focus-area":[],"msr-platform":[],"msr-download-source":[],"msr-locale":[268875],"msr-field-of-study":[246694,246691,246685,246901,250258,254305],"msr-conference":[],"msr-journal":[],"msr-impact-theme":[],"msr-pillar":[],"class_list":["post-663714","msr-research-item","type-msr-research-item","status-publish","hentry","msr-research-area-artificial-intelligence","msr-research-area-systems-and-networking","msr-locale-en_us","msr-field-of-study-artificial-intelligence","msr-field-of-study-computer-science","msr-field-of-study-machine-learning","msr-field-of-study-normalization-statistics","msr-field-of-study-skew","msr-field-of-study-skewed-data"],"msr_publishername":"PMLR","msr_edition":"","msr_affiliation":"","msr_published_date":"2020-7-11","msr_host":"","msr_duration":"","msr_version":"","msr_speaker":"","msr_other_contributors":"","msr_booktitle":"","msr_pages_string":"","msr_chapter":"","msr_isbn":"","msr_journal":"","msr_volume":"","msr_number":"","msr_editors":"","msr_series":"","msr_issue":"","msr_organization":"","msr_how_published":"","msr_notes":"","msr_highlight_text":"","msr_release_tracker_id":"","msr_original_fields_of_study":"","msr_download_urls":"","msr_external_url":"","msr_secondary_video_url":"","msr_longbiography":"","msr_microsoftintellectualproperty":0,"msr_main_download":"","msr_publicationurl":"","msr_doi":"","msr_publication_uploader":[{"type":"url","viewUrl":"false","id":"false","title":"https:\/\/www.microsoft.com\/en-us\/research\/uploads\/prod\/2020\/08\/Noniid_Camera.pdf","label_id":"243132","label":0},{"type":"url","viewUrl":"false","id":"false","title":"http:\/\/proceedings.mlr.press\/v119\/hsieh20a.html","label_id":"243109","label":0}],"msr_related_uploader":[{"type":"url","viewUrl":"false","id":"false","title":"https:\/\/icml.cc\/virtual\/2020\/poster\/6306","label_id":"243115","label":0},{"type":"url","viewUrl":"false","id":"false","title":"https:\/\/github.com\/kevinhsieh\/non_iid_dml","label_id":"243118","label":0}],"msr_attachments":[],"msr-author-ordering":[{"type":"user_nicename","value":"Kevin Hsieh","user_id":39459,"rest_url":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/microsoft-research\/v1\/researchers?person=Kevin Hsieh"},{"type":"user_nicename","value":"Amar Phanishayee","user_id":30975,"rest_url":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/microsoft-research\/v1\/researchers?person=Amar Phanishayee"},{"type":"user_nicename","value":"Onur Mutlu","user_id":33165,"rest_url":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/microsoft-research\/v1\/researchers?person=Onur Mutlu"},{"type":"text","value":"Phillip Gibbons","user_id":0,"rest_url":false}],"msr_impact_theme":[],"msr_research_lab":[199565],"msr_event":[670011],"msr_group":[144899,144927],"msr_project":[472845],"publication":[],"video":[],"download":[],"msr_publication_type":"inproceedings","_links":{"self":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item\/663714"}],"collection":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item"}],"about":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/types\/msr-research-item"}],"version-history":[{"count":5,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item\/663714\/revisions"}],"predecessor-version":[{"id":735010,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item\/663714\/revisions\/735010"}],"wp:attachment":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/media?parent=663714"}],"wp:term":[{"taxonomy":"msr-content-type","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-content-type?post=663714"},{"taxonomy":"msr-research-highlight","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-research-highlight?post=663714"},{"taxonomy":"msr-research-area","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/research-area?post=663714"},{"taxonomy":"msr-publication-type","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-publication-type?post=663714"},{"taxonomy":"msr-product-type","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-product-type?post=663714"},{"taxonomy":"msr-focus-area","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-focus-area?post=663714"},{"taxonomy":"msr-platform","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-platform?post=663714"},{"taxonomy":"msr-download-source","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-download-source?post=663714"},{"taxonomy":"msr-locale","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-locale?post=663714"},{"taxonomy":"msr-field-of-study","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-field-of-study?post=663714"},{"taxonomy":"msr-conference","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-conference?post=663714"},{"taxonomy":"msr-journal","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-journal?post=663714"},{"taxonomy":"msr-impact-theme","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-impact-theme?post=663714"},{"taxonomy":"msr-pillar","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-pillar?post=663714"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}