BEGIN:VCALENDAR VERSION:2.0 PRODID:IEEE vTools.Events//EN CALSCALE:GREGORIAN BEGIN:VTIMEZONE TZID:Asia/Kolkata BEGIN:STANDARD DTSTART:19451014T230000 TZOFFSETFROM:+0630 TZOFFSETTO:+0530 TZNAME:IST END:STANDARD END:VTIMEZONE BEGIN:VEVENT DTSTAMP:20230721T064617Z UID:E13239C2-C2BF-4F70-BDC6-04A85F2F56AE DTSTART;TZID=Asia/Kolkata:20230203T193000 DTEND;TZID=Asia/Kolkata:20230203T203000 DESCRIPTION:There is growing interest in solving large-scale statistical ma chine learning problems over decentralized networks\; where data are distr ibuted across the nodes of the network and no centralized coordination is present (we termed these systems “mesh” networks). Modern massive data sets create a fundamental problem at the intersection of the computational and statistical sciences: how to provide guarantees on the quality of sta tistical inference given bounds on computational resources\, such as time and communication efforts? While statistical-computation tradeoffs have be en largely explored in the centralized setting\, our understanding over me sh networks is limited: (i) distributed schemes\, designed and performing well in the classical low-dimensional regime\, can break down in the high- dimensional case\; and (ii) existing convergence studies may fail to predi ct algorithmic behaviors\; some are in fact confuted by experiments. This is mainly due to the fact that the majority of distributed algorithms have been designed and studied only from the optimization perspective\, lackin g the statistical dimension. This talk will discuss some vignettes from hi gh-dimensional statistical inference suggesting new analyses aiming at bri nging statistical thinking in distributed optimization.\n\nSpeaker(s): Dr. Gesualdo Scutari\, \n\nVirtual: https://events.vtools.ieee.org/m/346176 LOCATION:Virtual: https://events.vtools.ieee.org/m/346176 ORGANIZER:ieee.sps.sb.iitkgp@gmail.com SEQUENCE:4 SUMMARY:IEEE SPS SBC Webinar: Bringing Statistical Thinking in Distributed Optimization. Vignettes from statistical inference over Networks (By Dr. G esualdo Scutari) URL;VALUE=URI:https://events.vtools.ieee.org/m/346176 X-ALT-DESC:Description: <br /><p><span style="font-weight: 400\;">There is growing interest in solving large-scale statistical machine learning probl ems over decentralized networks\; where data are distributed across the no des of the network and no centralized coordination is present (we termed t hese systems &ldquo\;mesh&rdquo\; networks). Modern massive datasets creat e a fundamental problem at the intersection of the computational and stati stical sciences: how to provide guarantees on the quality of statistical i nference given bounds on computational resources\, such as time and commun ication efforts? While statistical-computation tradeoffs have been largely explored in the centralized setting\, our understanding over mesh network s is limited: (i) distributed schemes\, designed and performing well in th e classical low-dimensional regime\, can break down in the high-dimensiona l case\; and (ii) existing convergence studies may fail to predict algorit hmic behaviors\; some are in fact confuted by experiments. This is mainly due to the fact that the majority of distributed algorithms have been desi gned and studied only from the optimization perspective\, lacking the stat istical dimension. This talk will discuss some vignettes from high-dimensi onal statistical inference suggesting new analyses aiming at bringing stat istical thinking in distributed optimization.&nbsp\;</span></p> END:VEVENT END:VCALENDAR