SoftDICE for Imitation Learning: Rethinking Off-policy Distribution Matching
This addresses sample efficiency and stability issues in imitation learning for robotics and control applications, representing an incremental improvement over prior methods.
The paper tackles the problem of biased gradient estimates and unstable training in off-policy distribution matching for imitation learning by introducing SoftDICE, which recovers expert policies with only one demonstration trajectory and outperforms baselines on Mujoco tasks.
We present SoftDICE, which achieves state-of-the-art performance for imitation learning. SoftDICE fixes several key problems in ValueDICE, an off-policy distribution matching approach for sample-efficient imitation learning. Specifically, the objective of ValueDICE contains logarithms and exponentials of expectations, for which the mini-batch gradient estimate is always biased. Second, ValueDICE regularizes the objective with replay buffer samples when expert demonstrations are limited in number, which however changes the original distribution matching problem. Third, the re-parametrization trick used to derive the off-policy objective relies on an implicit assumption that rarely holds in training. We leverage a novel formulation of distribution matching and consider an entropy-regularized off-policy objective, which yields a completely offline algorithm called SoftDICE. Our empirical results show that SoftDICE recovers the expert policy with only one demonstration trajectory and no further on-policy/off-policy samples. SoftDICE also stably outperforms ValueDICE and other baselines in terms of sample efficiency on Mujoco benchmark tasks.