{"643579":{"#nid":"643579","#data":{"type":"event","title":"CSE Seminar with University of California, Berkeley Department of Statistics Post Doc Micha\u0142 Derezi\u0144ski ","body":[{"value":"\u003Cp\u003E\u003Cstrong\u003EName:\u003C\/strong\u003E\u0026nbsp;Micha\u0142 Derezi\u0144ski\u003C\/p\u003E\r\n\r\n\u003Cp\u003E\u003Cstrong\u003EDate\/Time:\u0026nbsp;\u003C\/strong\u003ETuesday, February 9 @ 11:00 am\u003C\/p\u003E\r\n\r\n\u003Cp\u003E\u003Cstrong\u003ELink:\u0026nbsp;\u003C\/strong\u003E\u003Ca href=\u0022https:\/\/nam12.safelinks.protection.outlook.com\/?url=https%3A%2F%2Fbluejeans.com%2F6622130444\u0026amp;data=04%7C01%7Ckristen.perez%40cc.gatech.edu%7C0f21edb0c65040f4dd4708d8c3b6b65b%7C482198bbae7b4b258b7a6d7f32faa083%7C0%7C0%7C637474536796501450%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000\u0026amp;sdata=hJ8d%2FXhZOmvIkQZADh3FDnxmXDCqmNClep6ECNNhx04%3D\u0026amp;reserved=0\u0022 title=\u0022\/\/bluejeans.com\/6622130444\r\n\r\nClick to follow link.\u0022\u003Ehttps:\/\/bluejeans.com\/6622130444\u003C\/a\u003E\u003C\/p\u003E\r\n\r\n\u003Cp\u003E\u003Cstrong\u003ETitle:\u003C\/strong\u003E\u0026nbsp;Bridging algorithmic and statistical randomness in machine learning\u003Cbr \/\u003E\r\n\u003Cbr \/\u003E\r\n\u003Cstrong\u003EAbstract:\u003C\/strong\u003E\u003Cbr \/\u003E\r\nRandomness is a key resource in designing efficient algorithms, and it is also a fundamental modeling framework in statistics and machine learning. Methods that lie at\u0026nbsp;the intersection of algorithmic and statistical randomness are at the forefront of modern data science. In this talk, I will discuss how statistical assumptions affect the bias-variance trade-offs and performance characteristics of randomized algorithms for, among others, linear regression, stochastic optimization, and dimensionality reduction. I\u0026nbsp;will also present an efficient algorithmic framework, called joint sampling, which is used to both predict and improve the statistical performance of machine learning\u0026nbsp;methods, by injecting carefully chosen correlations into randomized algorithms.\u003Cbr \/\u003E\r\n\u003Cbr \/\u003E\r\nIn the first part of the talk, I will focus on the phenomenon of inversion bias, which is a systematic bias caused by inverting random matrices. Inversion bias is a significant\u0026nbsp;bottleneck in parallel and distributed approaches to linear regression, second order optimization, and a range of statistical estimation tasks. Here, I will introduce a joint\u0026nbsp;sampling technique called Volume Sampling, which is the first method to eliminate inversion bias in model averaging. In the second part, I will demonstrate how the\u0026nbsp;spectral properties of data distributions determine the statistical performance of machine learning algorithms, going beyond worst-case analysis and revealing new phase\u0026nbsp;transitions in statistical learning. Along the way, I will highlight a class of joint sampling methods called Determinantal Point Processes (DPPs), popularized in machine\u0026nbsp;learning over the past fifteen years as a tractable model of diversity. In particular, I will present a new algorithmic technique called Distortion-Free Intermediate Sampling,\u0026nbsp;which drastically reduced the computational cost of DPPs, turning them into a practical tool for large-scale data science.\u0026nbsp;\u003Cbr \/\u003E\r\n\u003Cbr \/\u003E\r\n\u003Cstrong\u003EBio:\u003C\/strong\u003E\u003Cbr \/\u003E\r\nMicha\u0142 Derezi\u0144ski is a postdoctoral fellow in the Department of Statistics at the University of California, Berkeley. Previously, he was a research fellow at the Simons\u0026nbsp;Institute for the Theory of Computing (Fall 2018, Foundations of Data Science program). He obtained his Ph.D. in Computer Science at the University of California, Santa\u0026nbsp;Cruz, advised by professor Manfred Warmuth, where he received the Best Dissertation Award for his work on sampling methods in statistical learning. Micha\u0142\u0026#39;s current\u0026nbsp;research is focused on developing scalable randomized algorithms with robust statistical guarantees for machine learning, data science and optimization. His work on\u0026nbsp;reducing the cost of interpretability in dimensionality reduction received the Best Paper Award at the Thirty-fourth Conference on Neural Information Processing Systems.\u0026nbsp;More information is available at:\u0026nbsp;\u003Ca href=\u0022https:\/\/nam12.safelinks.protection.outlook.com\/?url=https:%2F%2Fusers.soe.ucsc.edu%2F~mderezin%2F\u0026amp;data=04%7C01%7Ckristen.perez%40cc.gatech.edu%7C0f21edb0c65040f4dd4708d8c3b6b65b%7C482198bbae7b4b258b7a6d7f32faa083%7C0%7C0%7C637474536796511437%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000\u0026amp;sdata=xhB68AdmCE13O5NY76NCu01LkLUsiWCPp3i18cAK7DA%3D\u0026amp;reserved=0\u0022\u003Ehttps:\/\/users.soe.ucsc.edu\/~mderezin\/\u003C\/a\u003E.\u003C\/p\u003E\r\n","summary":null,"format":"limited_html"}],"field_subtitle":"","field_summary":"","field_summary_sentence":[{"value":"CSE Seminar with University of California, Berkeley Department of Statistics Post Doc Micha\u0142 Derezi\u0144ski "}],"uid":"34540","created_gmt":"2021-01-28 19:18:37","changed_gmt":"2021-01-28 19:18:50","author":"Kristen Perez","boilerplate_text":"","field_publication":"","field_article_url":"","field_event_time":{"event_time_start":"2021-02-09T11:00:00-05:00","event_time_end":"2021-02-09T12:00:00-05:00","event_time_end_last":"2021-02-09T12:00:00-05:00","gmt_time_start":"2021-02-09 16:00:00","gmt_time_end":"2021-02-09 17:00:00","gmt_time_end_last":"2021-02-09 17:00:00","rrule":null,"timezone":"America\/New_York"},"extras":[],"groups":[{"id":"47223","name":"College of Computing"},{"id":"50877","name":"School of Computational Science and Engineering"}],"categories":[],"keywords":[],"core_research_areas":[],"news_room_topics":[],"event_categories":[{"id":"1795","name":"Seminar\/Lecture\/Colloquium"}],"invited_audience":[{"id":"78761","name":"Faculty\/Staff"},{"id":"78771","name":"Public"},{"id":"78751","name":"Undergraduate students"}],"affiliations":[],"classification":[],"areas_of_expertise":[],"news_and_recent_appearances":[],"phone":[],"contact":[{"value":"\u003Cp\u003EKristen Perez\u003C\/p\u003E\r\n\r\n\u003Cp\u003Ekristen.perez@cc.gatech.edu\u003C\/p\u003E\r\n","format":"limited_html"}],"email":[],"slides":[],"orientation":[],"userdata":""}}}