As the state of reinforcement learning has advanced, we wish to improve the performance and usability of the ALL in order to maintain it as a state-of-the-art reinforcement learning library.