BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Monte Carlo without rejection - Alexandre Bouchard (University of 
 British Columbia)
DTSTART:20170703T131500Z
DTEND:20170703T140000Z
UID:TALK73130@talks.cam.ac.uk
CONTACT:INI IT
DESCRIPTION:<span>Co-authors: Arnaud Doucet		(Oxford)\, Sebastian Vollmer	
 	(Warwick)\, George Deligiannidis		(King&#39\;s College London)\, Paul Van
 etti		(Oxford)        <br></span><span><br>Markov chain Monte Carlo method
 s have become standard tools to sample from complex high-dimensional proba
 bility measures. Many available techniques rely on discrete-time reversibl
 e Markov chains whose transition kernels built up over the Metropolis-Hast
 ings algorithm. In our recent work\, we investigate an alternative approac
 h\, the Bouncy Particle Sampler (BPS) where the target distribution of int
 erest is explored using a continuous-time\, non reversible Markov process.
  In this alternative approach\, a particle moves along straight lines cont
 inuously around the space and\, when facing a high energy barrier\, it is 
 not rejected but its path is modified by bouncing against this barrier. Th
 e resulting non-reversible Markov process provides a rejection-free Markov
  chain Monte Carlo sampling scheme. This method\, inspired from recent wor
 k in the molecular simulation literature\, is shown to be a valid\, effici
 ent sampling scheme applicable to a wide range of Bayesian problems. We pr
 esent  several additional original methodological extensions and establish
  various theoretical properties of these procedures. We demonstrate experi
 mentally the efficiency of these algorithms on a variety of Bayesian infer
 ence problems.</span>
LOCATION:Seminar Room 1\, Newton Institute
END:VEVENT
END:VCALENDAR