Diffusion Model-Augmented Behavioral Cloning
This work addresses imitation learning for robotics and control applications, offering a method that balances generalization and efficiency, though it is incremental by building on existing behavioral cloning and diffusion model approaches.
The paper tackles the generalization and inference efficiency challenges in imitation learning by proposing Diffusion Model-Augmented Behavioral Cloning (DBC), which combines conditional and joint probability modeling of expert distributions, resulting in improved performance over baselines in continuous control tasks such as navigation, robot arm manipulation, dexterous manipulation, and locomotion.
Imitation learning addresses the challenge of learning by observing an expert's demonstrations without access to reward signals from environments. Most existing imitation learning methods that do not require interacting with environments either model the expert distribution as the conditional probability p(a|s) (e.g., behavioral cloning, BC) or the joint probability p(s, a). Despite the simplicity of modeling the conditional probability with BC, it usually struggles with generalization. While modeling the joint probability can improve generalization performance, the inference procedure is often time-consuming, and the model can suffer from manifold overfitting. This work proposes an imitation learning framework that benefits from modeling both the conditional and joint probability of the expert distribution. Our proposed Diffusion Model-Augmented Behavioral Cloning (DBC) employs a diffusion model trained to model expert behaviors and learns a policy to optimize both the BC loss (conditional) and our proposed diffusion model loss (joint). DBC outperforms baselines in various continuous control tasks in navigation, robot arm manipulation, dexterous manipulation, and locomotion. We design additional experiments to verify the limitations of modeling either the conditional probability or the joint probability of the expert distribution, as well as compare different generative models. Ablation studies justify the effectiveness of our design choices.