Primal Wasserstein Imitation Learning
This work addresses imitation learning for robotics and control systems, offering a simpler and more efficient alternative to adversarial methods, though it appears incremental as it builds on existing Wasserstein distance concepts.
The authors tackled the problem of imitation learning by proposing Primal Wasserstein Imitation Learning (PWIL), a method that uses the primal Wasserstein distance to match expert behavior, achieving sample-efficient recovery of expert behavior on continuous control tasks in MuJoCo.
Imitation Learning (IL) methods seek to match the behavior of an agent with that of an expert. In the present work, we propose a new IL method based on a conceptually simple algorithm: Primal Wasserstein Imitation Learning (PWIL), which ties to the primal form of the Wasserstein distance between the expert and the agent state-action distributions. We present a reward function which is derived offline, as opposed to recent adversarial IL algorithms that learn a reward function through interactions with the environment, and which requires little fine-tuning. We show that we can recover expert behavior on a variety of continuous control tasks of the MuJoCo domain in a sample efficient manner in terms of agent interactions and of expert interactions with the environment. Finally, we show that the behavior of the agent we train matches the behavior of the expert with the Wasserstein distance, rather than the commonly used proxy of performance.