Human-level performance in first-person multiplayer games with population-based deep reinforcement learning | Read Paper on Bytez