Policy-shaped prediction: avoiding distractions in model-based reinforcement learning

Re-assign community