And it’s a core part of Google subsidiary DeepMind’s deep Q ... trained using reinforcement learning. To that end, it includes a compact set of well-documented code (15 Python files) focused ...