BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Variational Bayes for high-dimensional linear regression with spar
 se priors - Kolyan Ray (Imperial College London)
DTSTART:20220223T140000Z
DTEND:20220223T150000Z
UID:TALK170651@talks.cam.ac.uk
CONTACT:Randolf Altmeyer
DESCRIPTION:A core problem in Bayesian statistics is approximating difficu
 lt to compute posterior distributions. In variational Bayes (VB)\, a metho
 d from machine learning\, one approximates the posterior through optimizat
 ion\, which is typically faster than Markov chain Monte Carlo. We study a 
 mean-field (i.e. factorizable) VB approximation to Bayesian model selectio
 n priors\, including the popular spike-and-slab prior\, in sparse high-dim
 ensional linear regression. We establish convergence rates for this VB app
 roach\, studying conditions under which it provides good estimation. We al
 so discuss some computational issues and study the empirical performance o
 f the algorithm.
LOCATION:MR3\, Centre for Mathematical Sciences
END:VEVENT
END:VCALENDAR