{"60899":{"#nid":"60899","#data":{"type":"event","title":"CSE Seminar: Santosh Vempala","body":[{"value":"\u003Cp\u003E\u003Cstrong\u003ESpeaker:\u0026nbsp;Santosh Vempala (Georgia Tech)\u003C\/strong\u003E\u003C\/p\u003E\u003Cp\u003E\u003Cstrong\u003E\u0022The Joy of PCA\u0022\u003C\/strong\u003E\u003C\/p\u003E\u003Cp\u003E\u003Cstrong\u003EAbstract:\u003C\/strong\u003E\u003C\/p\u003E\u003Cp\u003EPrincipal\n Component Analysis is the most widely used technique for \nhigh-dimensional or large data. For typical applications (nearest \nneighbor, clustering, learning), it is not hard to build examples on \nwhich PCA *fails*. Yet, it is popular and successful across a variety of\n data-rich areas. In this talk, we focus on two algorithmic problems \nwhere the performance of PCA is provably near-optimal, and no other \nmethod is known to have similar guarantees. The problems we consider are\n (a) the classical statistical problem of unraveling a sample from a \nmixture of k unknown Gaussians and (b) the classic learning theory \nproblem of learning an intersection of k halfspaces. During the talk, we\n will encounter recent extensions of PCA that are noise-resistant, \naffine-invariant and nonviolent.\u003C\/p\u003E","summary":null,"format":"limited_html"}],"field_subtitle":"","field_summary":"","field_summary_sentence":[{"value":"\u0022The Joy of PCA\u0022"}],"uid":"27345","created_gmt":"2010-09-10 10:19:54","changed_gmt":"2016-10-08 01:52:18","author":"Cristina Gonzalez","boilerplate_text":"","field_publication":"","field_article_url":"","field_event_time":{"event_time_start":"2010-09-17T03:00:00-04:00","event_time_end":"2010-09-17T03:00:00-04:00","event_time_end_last":"2010-09-17T03:00:00-04:00","gmt_time_start":"2010-09-17 07:00:00","gmt_time_end":"2010-09-17 07:00:00","gmt_time_end_last":"2010-09-17 07:00:00","rrule":null,"timezone":"America\/New_York"},"extras":[],"groups":[{"id":"47223","name":"College of Computing"}],"categories":[],"keywords":[{"id":"3497","name":"cse seminar"},{"id":"167405","name":"santosh vempala"}],"core_research_areas":[],"news_room_topics":[],"event_categories":[{"id":"1795","name":"Seminar\/Lecture\/Colloquium"}],"invited_audience":[],"affiliations":[],"classification":[],"areas_of_expertise":[],"news_and_recent_appearances":[],"phone":[],"contact":[],"email":[],"slides":[],"orientation":[],"userdata":""}}}