{"id":188630,"date":"2012-10-26T00:00:00","date_gmt":"2012-10-26T15:58:23","guid":{"rendered":"https:\/\/www.microsoft.com\/en-us\/research\/msr-research-item\/machine-learning-work-shop-session-2-ofer-dekel-online-learning-against-adaptive-adversaries\/"},"modified":"2016-08-22T11:27:13","modified_gmt":"2016-08-22T18:27:13","slug":"machine-learning-work-shop-session-2-ofer-dekel-online-learning-against-adaptive-adversaries","status":"publish","type":"msr-video","link":"https:\/\/www.microsoft.com\/en-us\/research\/video\/machine-learning-work-shop-session-2-ofer-dekel-online-learning-against-adaptive-adversaries\/","title":{"rendered":"Machine Learning Work Shop &#8211; Session 2 &#8211; Ofer Dekel &#8211; &#8220;Online Learning Against Adaptive Adversaries&#8221;"},"content":{"rendered":"<div class=\"asset-content\">\n<p>Most machine learning algorithms rely on the assumption that the data is generated by a stochastic process. Many online learning algorithms go one step further and allow the data to be generated by an oblivious adversary (a.k.a. a non-adaptive or non-reactive adversary). Very little is known about the machine learning scenario where the data is generated by a more powerful adversary, such as a switching-cost adversary, a memory-bounded adversary, or a general adaptive adversary. In this talk, I will describe ongoing efforts to understand what can and cannot be done against these powerful rivals. First, I will define policy-regret, which is a meaningful way of measuring the performance of a learning algorithm in the adversarial setting. Then, I will present the current state-of-the-art upper and lower bounds on policy regret against different adversary types, in the full-information and the bandit-feedback settings.<\/p>\n<p>This talk represents joint work with Ambuj Tewari, Raman Arora, Nicolo Cesa-Bianchi, and Ohad Shamir.<\/p>\n<\/div>\n<p><!-- .asset-content --><\/p>\n","protected":false},"excerpt":{"rendered":"<p>Most machine learning algorithms rely on the assumption that the data is generated by a stochastic process. Many online learning algorithms go one step further and allow the data to be generated by an oblivious adversary (a.k.a. a non-adaptive or non-reactive adversary). Very little is known about the machine learning scenario where the data is [&hellip;]<\/p>\n","protected":false},"featured_media":197262,"template":"","meta":{"msr-url-field":"","msr-podcast-episode":"","msrModifiedDate":"","msrModifiedDateEnabled":false,"ep_exclude_from_search":false,"_classifai_error":"","msr_hide_image_in_river":0,"footnotes":""},"research-area":[],"msr-video-type":[],"msr-locale":[268875],"msr-post-option":[],"msr-session-type":[],"msr-impact-theme":[],"msr-pillar":[],"msr-episode":[],"msr-research-theme":[],"class_list":["post-188630","msr-video","type-msr-video","status-publish","has-post-thumbnail","hentry","msr-locale-en_us"],"msr_download_urls":"","msr_external_url":"https:\/\/youtu.be\/9mWRzgKazi8","msr_secondary_video_url":"","msr_video_file":"","_links":{"self":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-video\/188630","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-video"}],"about":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/types\/msr-video"}],"version-history":[{"count":0,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-video\/188630\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/media\/197262"}],"wp:attachment":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/media?parent=188630"}],"wp:term":[{"taxonomy":"msr-research-area","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/research-area?post=188630"},{"taxonomy":"msr-video-type","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-video-type?post=188630"},{"taxonomy":"msr-locale","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-locale?post=188630"},{"taxonomy":"msr-post-option","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-post-option?post=188630"},{"taxonomy":"msr-session-type","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-session-type?post=188630"},{"taxonomy":"msr-impact-theme","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-impact-theme?post=188630"},{"taxonomy":"msr-pillar","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-pillar?post=188630"},{"taxonomy":"msr-episode","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-episode?post=188630"},{"taxonomy":"msr-research-theme","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-research-theme?post=188630"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}