Cross-Domain Imitation Learning via Optimal Transport
This addresses the challenge of leveraging expert demonstrations for agents with different morphologies, which is incremental as it builds on optimal transport methods.
The paper tackles the problem of cross-domain imitation learning by proposing GWIL, which uses the Gromov-Wasserstein distance to align states between agents with different embodiments, and demonstrates its effectiveness in continuous control domains with transformations of state-action spaces.
Cross-domain imitation learning studies how to leverage expert demonstrations of one agent to train an imitation agent with a different embodiment or morphology. Comparing trajectories and stationary distributions between the expert and imitation agents is challenging because they live on different systems that may not even have the same dimensionality. We propose Gromov-Wasserstein Imitation Learning (GWIL), a method for cross-domain imitation that uses the Gromov-Wasserstein distance to align and compare states between the different spaces of the agents. Our theory formally characterizes the scenarios where GWIL preserves optimality, revealing its possibilities and limitations. We demonstrate the effectiveness of GWIL in non-trivial continuous control domains ranging from simple rigid transformation of the expert domain to arbitrary transformation of the state-action space.