Bibliography (3):

1. Human-level performance in 3D multiplayer games with population-based reinforcement learning

2. A general reinforcement learning algorithm that masters chess, shogi, and Go through self-play

3. https://github.com/pytorch/ELF