Modular Lifelong Reinforcement Learning via Neural Composition.

Jorge A. Mendez, Harm van Seijen, Eric Eaton

Browse the full ICLR paper archive.