BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:Solving the Poisson equation using coupled Markov
DTSTART:20221202T151500
DTEND:20221202T170000
DTSTAMP:20260410T111602Z
UID:b414e300db0729190f2be18df83744b7a9053277d6dfc893ce827f77
CATEGORIES:Conferences - Seminars
DESCRIPTION:Pierre Jacob (ESSEC Business School\, Paris)\nThis work starts
  with connections between couplings of Markov chains and solutions of the 
 Poisson equation. We show how pairs of chains can be employed to obtain un
 biased estimators of pointwise evaluations of solutions of the Poisson equ
 ation. Motivated by MCMC\, we propose new estimators of the asymptotic var
 iance of Markov chain ergodic averages. The proposed estimators have disti
 nct appeals and drawbacks relative to standard methods\, such as batch mea
 ns or spectral variance methods. We formally study the proposed estimators
  under realistic assumptions on the meeting times of the coupled chains an
 d on the existence of moments of test functions under the target distribut
 ion. We describe experiments in toy examples and more challenging settings
  arising in Bayesian analysis. This is joint work with Randal Douc\, Antho
 ny Lee and Dootika Vats.\n 
LOCATION:MA A1 12 https://plan.epfl.ch/?room==MA%20A1%2012
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR