Offline Imitation Learning by Controlling the Effective Planning Horizon
This addresses a specific issue in offline imitation learning for robotics or AI agents, but it is incremental as it builds on existing methods by correcting errors rather than introducing a new paradigm.
The paper tackles the problem of sampling errors in offline imitation learning by controlling the effective planning horizon instead of using explicit regularization, showing that the corrected algorithm improves performance on popular benchmarks.
In offline imitation learning (IL), we generally assume only a handful of expert trajectories and a supplementary offline dataset from suboptimal behaviors to learn the expert policy. While it is now common to minimize the divergence between state-action visitation distributions so that the agent also considers the future consequences of an action, a sampling error in an offline dataset may lead to erroneous estimates of state-action visitations in the offline case. In this paper, we investigate the effect of controlling the effective planning horizon (i.e., reducing the discount factor) as opposed to imposing an explicit regularizer, as previously studied. Unfortunately, it turns out that the existing algorithms suffer from magnified approximation errors when the effective planning horizon is shortened, which results in a significant degradation in performance. We analyze the main cause of the problem and provide the right remedies to correct the algorithm. We show that the corrected algorithm improves on popular imitation learning benchmarks by controlling the effective planning horizon rather than an explicit regularization.