Learning to Plan Probabilistically from Neural Networks

| This paper discusses the learning of probabilis-tic planning without a priori domain-speciic knowledge. Diierent from existing reinforcement learning algorithms that generate only reactive policies and existing probabilis-tic planning algorithms that requires a substantial amount of a priori knowledge in order to plan, we devise a two-stage bottom-up learning-to-plan process, in which rst reinforcement learning/dynamic programming is applied, without the use of a priori domain-speciic knowledge, to acquire a reactive policy and then explicit plans are extracted from the learned reactive policy. Plan extraction is based on a beam search algorithm that performs temporal projection in a restricted fashion guided by the value functions resulting from reinforcement learning/dynamic programming. Experiments and theoretical analysis are presented.