This “Imitation” Python library provides open-source implementations of imitation and reward learning algorithms in PyTorch
In domains where reward functions are clearly defined, such as games, reinforcement learning (RL) has outperformed human performance. Unfortunately, it is difficult or impossible for many real-world tasks to design the reward function procedurally. Instead, they should immediately absorb a reward feature or policy from user feedback. Moreover, even when a reward function can be …