BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//MIDAS - ECPv5.1.5//NONSGML v1.0//EN
CALSCALE:GREGORIAN
METHOD:PUBLISH
X-WR-CALNAME:MIDAS
X-ORIGINAL-URL:https://midas.umich.edu
X-WR-CALDESC:Events for MIDAS
BEGIN:VTIMEZONE
TZID:America/Detroit
BEGIN:DAYLIGHT
TZOFFSETFROM:-0500
TZOFFSETTO:-0400
TZNAME:EDT
DTSTART:20190310T070000
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0400
TZOFFSETTO:-0500
TZNAME:EST
DTSTART:20191103T060000
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTART;TZID=America/Detroit:20190314T160000
DTEND;TZID=America/Detroit:20190314T170000
DTSTAMP:20210928T100424
CREATED:20190312T163629Z
LAST-MODIFIED:20190312T163629Z
UID:28460-1552579200-1552582800@midas.umich.edu
SUMMARY:CSP Seminar\, David Gamarnik (MIT)\, Thursday\, 3/14\, 4:00pm\, Room 1005 EECS
DESCRIPTION:\nCommunications and Signal Processing Seminar (CSP)\nDavid Gamarnik\, PhD\n\nProfessor of Operations Research at the Operations Research and Statistics Group\nSloan School of Management of Massachusetts Institute of Technology\n\n“Algorithms and Algorithmic Obstacles in High-Dimensional Regression“\n\nAbstract: Many optimization problems arising in studying of high-dimensional inference problems exhibit an apparent gap between the optimal values which can be estimated by non-constructive means\, and the best values achievable by fast (polynomial time) algorithms. Recently it became apparent that a potential and in some cases a provable obstruction for designing algorithms bridging this gap is a phase transition in the geometry of nearly optimal solutions\, in particular the presence of a certain Overlap Gap Property (OGP). In this talk we will discuss this property in the context of sparse high dimensional linear regression problem. We show that\, on the one hand\, in the sampling regime where the known fast methods for this problem are effective\, including LASSO\, Basis Pursuit\, Orthogonal Matching Pursuit\, the space of solutions exhibits a monotonicity with respect to the proximity to the ground truth regression vector and no local optimums exist apart from the ground truth. On the other hand\, once the sampling number is asymptotically in the regime where the known methods fail\, we show that the monotonicity is lost\, and the model exhibits an OGP. In the context of the regression problem this means every solution exhibiting a small mean squared error is either fairly close to the ground truth or is very far from it\, with no middle ground. Joint work with Ilias Zadik (MIT).\n\nBiography: David Gamarnik is a Professor of Operations Research at the Operations Research and Statistics Group\, Sloan School of Management of Massachusetts Institute of Technology. He received B.A. in mathematics from New York University in 1993 and Ph.D. in Operations Research from MIT in 1998. Since then he was a research staff member of IBM T.J. Watson Research Center\, before joining MIT in 2005. His research interests include probability\, theory of random graphs\, optimization and algorithms\, statistics and machine learning\, stochastic processes and queueing theory. He is a recipient of the Erlang Prize and the Best Publication Award from the Applied Probability Society of INFORMS\, and he was a finalist for Franz Edelman Prize competition of INFORMS. He serves and served on the editorial boards of several journals including Operations Research\, Mathematics of Operations Research\, Annals of Applied Probability\, Queueing Systems and Stochastic Systems. \n\nhttps://www.eecs.umich.edu/eecs/etc/events/showevent.cgi?5071\n\n
URL:https://midas.umich.edu/event/csp-seminar-david-gamarnik-mit-thursday-3-14-400pm-room-1005-eecs/
LOCATION:1005 EECS\, 1301 Beal Ave.\, Ann Arbor\, Michigan\, 48109
END:VEVENT
END:VCALENDAR