Zero-Shot Offline Imitation Learning via Optimal Transport
This addresses the issue of short-sighted actions undermining long-term objectives in imitation learning for robotics or AI agents, representing an incremental improvement over existing goal-based methods.
The paper tackles the problem of myopic behavior in zero-shot imitation learning by directly optimizing occupancy matching, enabling non-myopic imitation from offline, suboptimal data in complex, continuous benchmarks.
Zero-shot imitation learning algorithms hold the promise of reproducing unseen behavior from as little as a single demonstration at test time. Existing practical approaches view the expert demonstration as a sequence of goals, enabling imitation with a high-level goal selector, and a low-level goal-conditioned policy. However, this framework can suffer from myopic behavior: the agent's immediate actions towards achieving individual goals may undermine long-term objectives. We introduce a novel method that mitigates this issue by directly optimizing the occupancy matching objective that is intrinsic to imitation learning. We propose to lift a goal-conditioned value function to a distance between occupancies, which are in turn approximated via a learned world model. The resulting method can learn from offline, suboptimal data, and is capable of non-myopic, zero-shot imitation, as we demonstrate in complex, continuous benchmarks. The code is available at https://github.com/martius-lab/zilot.