A. Tavakoli, F. Pardo, and P. Kormushev, “Action Branching Architectures for Deep Reinforcement Learning”, AAAI, vol. 32, no. 1, Apr. 2018.