Thompson sampling for Markov games with piecewise stationary opponent policies.

Anthony DiGiovanni, Ambuj Tewari

Browse the full UAI paper archive.