BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Rich Component Analysis - James Zou (Microsoft Research New Englan
 d)
DTSTART:20150717T100000Z
DTEND:20150717T110000Z
UID:TALK59875@talks.cam.ac.uk
CONTACT:Dr Jes Frellsen
DESCRIPTION:In many settings\, we have multiple data sets (also called vie
 ws) that capture different and overlapping aspects of the same phenomenon.
  We are often interested in finding patterns that are unique to one or to 
 a subset of the views. For example\, we might have one set of molecular ob
 servations and one set of physiological observations on the same group of 
 individuals\, and we want to quantify molecular patterns that are uncorrel
 ated with physiology. Despite being a common problem\, this is highly chal
 lenging when the correlations come from complex distributions. In this pap
 er\, we develop the general framework of Rich Component Analysis (RCA) to 
 model settings where the observations from different views are driven by d
 ifferent sets of latent components\, and each component can be a complex\,
  high-dimensional distribution. We introduce algorithms based on cumulant 
 extraction that provably learn each of the components without having to mo
 del the other components. We show how to integrate RCA with stochastic gra
 dient descent into a meta-algorithm for learning general models\, and demo
 nstrate substantial improvement in accuracy on several synthetic and real 
 datasets in both supervised and unsupervised tasks.  Our method makes it p
 ossible to learn latent variable models when we don't have samples from th
 e true model but only samples after complex perturbations.\n\nBased on joi
 nt work with Rong Ge (MSR).
LOCATION:Engineering Department\, CBL Room BE-438
END:VEVENT
END:VCALENDAR