[YPNG] YPNG seminar 26 Jan 2018, in Stats classroom 11:00-1:00

[David Pollard]

Elena and David have been trying to understand some papers of Lezaud (1998 Ann. Appl. Prob)
and Paulin (2015 Elect. J Prob) on concentration for sums f(X_1)+ \dots + f(X_n), where
X_0,X_1,\dots is a time-reversible Markov chain with a finite state space. Their method involves a
lot of fancy perturbation theory, a la Kato. On the assumption that many in the audience will have
forgotten any complex analysis that they might have known, David will try to explain everything,
including the basics of spectral theory for Markov chains.
The talk will be very boring for anyone who already knows a lot about spectral theory and eigengaps.
It is intended to provide a gentle introduction for those who have never experienced a YPNG talk.