Improved Planning for Infinite-Horizon Interactive POMDPs Using Probabilistic Inference

We provide the first formalization of self-interested multiagent planning using expectation-maximization (EM). Our formalization in the context of infinite-horizon and finitely-nested interactive POMDP (I-POMDP) is distinct from EM formulations for POMDPs and other multiagent planning frameworks. Specific to I-POMDPs, we exploit the graphical model structure and present a new approach based on block-coordinate descent for further speed up.