BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//talks.cam.ac.uk//v3//EN
BEGIN:VTIMEZONE
TZID:Europe/London
BEGIN:DAYLIGHT
TZOFFSETFROM:+0000
TZOFFSETTO:+0100
TZNAME:BST
DTSTART:19700329T010000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=-1SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0100
TZOFFSETTO:+0000
TZNAME:GMT
DTSTART:19701025T020000
RRULE:FREQ=YEARLY;BYMONTH=10;BYDAY=-1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
CATEGORIES:Isaac Newton Institute Seminar Series
SUMMARY:A Tensor Spectral Approach to Learning Mixed Membe
 rship Community Models - Anandkumar \, A (Universi
 ty of California\, Irvine)
DTSTART;TZID=Europe/London:20130813T100000
DTEND;TZID=Europe/London:20130813T104500
UID:TALK46607AThttp://talks.cam.ac.uk
URL:http://talks.cam.ac.uk/talk/index/46607
DESCRIPTION:Co-authors: Rong Ge (Princeton)\, Daniel Hsu (Micr
 osoft Research)\, Sham Kakade (Microsoft Research)
  \n\nModeling community formation and detecting hi
 dden communities in networks is a well studied pro
 blem. However\, theoretical analysis of community 
 detection has been mostly limited to models with n
 on-overlapping communities such as the stochastic 
 block model. In this paper\, we remove this restri
 ction\, and consider a family of probabilistic net
 work models with overlapping communities\, termed 
 as the mixed membership Dirichlet model\, first in
 troduced in Aioroldi et. al. 2008. This model allo
 ws for nodes to have fractional memberships in mul
 tiple communities and assumes that the community m
 emberships are drawn from a Dirichlet distribution
 . We propose a unified approach to learning these 
 models via a tensor spectral decomposition method.
  Our estimator is based on low-order moment tensor
  of the observed network\, consisting of 3-star co
 unts. Our learning method is fast and is based on 
 simple linear algebra operations\, e.g. singular v
 alue decomposition and tensor power iterations. We
  provide guaranteed recovery of community membersh
 ips and model parameters and present a careful fin
 ite sample analysis of our learning method. Additi
 onally\, our results match the best known scaling 
 requirements in the special case of the stochastic
  block model.\n\nRelated Links 	http://newport.eec
 s.uci.edu/anandkumar/pubs/AnandkumarCommunity.pdf 
 - manuscript\n\n
LOCATION:Seminar Room 1\, Newton Institute
CONTACT:Mustapha Amrani
END:VEVENT
END:VCALENDAR