BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Coping with the Intractability of Graphical Models - Justin Domke
DTSTART:20141127T100000Z
DTEND:20141127T110000Z
UID:TALK56481@talks.cam.ac.uk
CONTACT:Microsoft Research Cambridge Talks Admins
DESCRIPTION:Many potential applications of graphical models (such as Condi
 tional Random Fields) are complicated by the fact that exact inference is 
 intractable.  This talk will describe two strategies for coping with this 
 situation.  The first is based on restricting consideration to a tractable
  set of parameters.  Rather than tree-structured parameters\, as is common
 \, I will explore a notion of tractability where Markov chain Monte Carlo 
 is guaranteed to quickly converge to the stationary distribution.  This ca
 n be used both for inference (as a type of generalized mean-field algorith
 m) and for learning\, where it gives a FPRAS for maximum likelihood learni
 ng when restricted to this set.  The second\, more pragmatic\, strategy is
  based on empirical risk minimization\, where a given approximate inferenc
 e method is “baked in” to the loss function.  In particular\, I will a
 lso discuss a recently released open-source tool for distributed learning 
 of such models using MPI.
LOCATION:Auditorium\, Microsoft Research Ltd\, 21 Station Road\, Cambridge
 \, CB1 2FB
END:VEVENT
END:VCALENDAR