DIDI: Diffusion-Guided Diversity for Offline Behavioral Generation
This work addresses the challenge of generating diverse behaviors from sub-optimal offline data for decision-making systems, representing an incremental improvement with novel method integration.
The paper tackles the problem of learning diverse skills from label-free offline data by proposing DIDI, which uses diffusion models as priors to guide and regularize policy learning, resulting in effective discovery of diverse and discriminative behaviors across multiple decision-making domains.
In this paper, we propose a novel approach called DIffusion-guided DIversity (DIDI) for offline behavioral generation. The goal of DIDI is to learn a diverse set of skills from a mixture of label-free offline data. We achieve this by leveraging diffusion probabilistic models as priors to guide the learning process and regularize the policy. By optimizing a joint objective that incorporates diversity and diffusion-guided regularization, we encourage the emergence of diverse behaviors while maintaining the similarity to the offline data. Experimental results in four decision-making domains (Push, Kitchen, Humanoid, and D4RL tasks) show that DIDI is effective in discovering diverse and discriminative skills. We also introduce skill stitching and skill interpolation, which highlight the generalist nature of the learned skill space. Further, by incorporating an extrinsic reward function, DIDI enables reward-guided behavior generation, facilitating the learning of diverse and optimal behaviors from sub-optimal data.