Learning Generalized Reactive Policies using Deep Neural Networks

A new approach to learning for planning, where knowledge acquired while solving a given set of planning problems is used to plan faster in related, but new problem instances.

BibTex: