Search papers, labs, and topics across Lattice.
This paper introduces Latent Wasserstein Adversarial Imitation Learning (LWAIL), a novel adversarial imitation learning framework designed to match state-only distributions using the Wasserstein distance in a dynamics-aware latent space. The dynamics-aware latent space is learned by pre-training an Intention Conditioned Value Function (ICVF) on randomly generated state data to capture state transition information. Experiments on MuJoCo environments demonstrate that LWAIL achieves expert-level performance using only one or a few state-only expert episodes, outperforming existing Wasserstein-based and adversarial IL methods.
Achieve expert-level imitation learning with just a handful of state-only expert demonstrations by pre-training a dynamics-aware latent space.
Imitation Learning (IL) enables agents to mimic expert behavior by learning from demonstrations. However, traditional IL methods require large amounts of medium-to-high-quality demonstrations as well as actions of expert demonstrations, both of which are often unavailable. To reduce this need, we propose Latent Wasserstein Adversarial Imitation Learning (LWAIL), a novel adversarial imitation learning framework that focuses on state-only distribution matching. It benefits from the Wasserstein distance computed in a dynamics-aware latent space. This dynamics-aware latent space differs from prior work and is obtained via a pre-training stage, where we train the Intention Conditioned Value Function (ICVF) to capture a dynamics-aware structure of the state space using a small set of randomly generated state-only data. We show that this enhances the policy's understanding of state transitions, enabling the learning process to use only one or a few state-only expert episodes to achieve expert-level performance. Through experiments on multiple MuJoCo environments, we demonstrate that our method outperforms prior Wasserstein-based IL methods and prior adversarial IL methods, achieving better results across various tasks.
