BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Community Detection on the Weighted Stochastic Block Model - Min X
 u (U Penn)
DTSTART:20161014T150000Z
DTEND:20161014T160000Z
UID:TALK67486@talks.cam.ac.uk
CONTACT:Quentin Berthet
DESCRIPTION:Ever since the seminal paper of Decelle et al appeared in 2011
 \, Stochastic Block Model has become the most well-studied and well-unders
 tood model for network data with an underlying community structure. Yet SB
 M has a limitation: it assumes that each network edge is Bernoulli 0/1--ei
 ther on or off\; this is restrictive because weighted edges are ubiquitous
  and\, when edge weights are present\, it may be important to incorporate 
 them into a clustering algorithm. In this talk\, we study the weighted gen
 eralization of the stochastic block model in which an edge random variable
  can have a general mixed distribution rather than Bernoulli\; we propose 
 and analyze an algorithm for the weighted SBM based a binning procedure fo
 r nonparametric density estimation. We show that this procedure has error 
 rate exponential in the information divergence that governs the thresholds
  for the unweighted Stochastic Block Model--a rate that in many cases have
  matching lower bounds. \n\nJoint work with Varun Jog and Po-Ling Loh from
  University of Wisconsin Madison and Zongming Ma from the University of Pe
 nnsylvania. 
LOCATION:MR12\, Centre for Mathematical Sciences\, Wilberforce Road\, Camb
 ridge.
END:VEVENT
END:VCALENDAR