BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Bayes-optimal estimation in generalized linear models - Ramji Venk
 ataramanan\, Department of Engineering\, Cambridge
DTSTART:20230215T140000Z
DTEND:20230215T150000Z
UID:TALK197254@talks.cam.ac.uk
CONTACT:Prof. Ramji Venkataramanan
DESCRIPTION:We consider the problem of signal estimation in  generalized l
 inear models (GLM)\, a class of models which includes canonical problems s
 uch as linear regression\, logistic regression\, and phase retrieval. Rece
 nt work has precisely characterized the asymptotic minimum mean-squared er
 ror (MMSE)  for  GLMs  with i.i.d. Gaussian measurement matrices. However\
 , in many models there is a significant gap between the MMSE and the perfo
 rmance of the best known feasible estimators. To address this\, we conside
 r GLMs defined via _spatially coupled_ measurement matrices. We propose an
  efficient approximate message passing (AMP) algorithm for estimation and 
 prove that the error of a carefully tuned AMP estimator approaches the asy
 mptotic MMSE. \n\nThe talk will not assume any background on message passi
 ng or spatial coupling. Joint work with Pablo Pascual Cobo and Kuan Hsieh.
LOCATION:MR5\, CMS Pavilion A
END:VEVENT
END:VCALENDAR