BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Quantitative convergence bounds for unadjusted kinetic Langevin an
 d Hamiltonian Monte Carlo - Pierre Monmarché (Sorbonne Université)
DTSTART:20241128T112500Z
DTEND:20241128T121500Z
UID:TALK221551@talks.cam.ac.uk
DESCRIPTION:Splitting schemes for the Hamiltonian and (underdamped) Langev
 in dynamics\, which are kinetic (possibly non-reversible) processes\, are 
 widely used in Markov Chain Monte Carlo methods. We will discuss how\, for
  Gaussian distributions\, a very precise optimization of the parameters ca
 n be conducted\, revealing how inertia induces a diffusive-to-ballistic sp
 eed-up for ill-conditioned targets. Motivated by this\, we will present no
 n-asymptotic efficiency bounds for this family of MCMC samplers that cover
  non-convex potentials and mean-field interacting particles.
LOCATION:Seminar Room 1\, Newton Institute
END:VEVENT
END:VCALENDAR