Self-Guided Action Diffusion

Rhea Malhotra, Yuejiang Liu, Chelsea Finn

Keywords: Inference-time Adaptation, Trajectory Coherence, Temporal Consistency, Policy Generalization

TL;DR: Self-GAD is an inference-time guidance method that improves diffusion-based policies by applying prior-conditioned gradient updates to enhance performance, robustness, and sample efficiency in dynamic, diverse, and out-of-distribution settings.

Abstract: Recent works have shown the promise of inference-time search over action samples for improving generative robot policies. In particular, optimizing cross-chunk coherence via bidirectional decoding has proven effective in boosting the consistency and reactivity of diffusion policies. However, this approach remains computationally expensive as the diversity of sampled actions grows. In this paper, we introduce self-guided action diffusion, a more efficient variant of bidirectional decoding tailored for diffusion-based policies. At the core of our method is to guide the proposal distribution at each diffusion step based on the prior decision. Experiments on robotic foundation model GR00T-N1 and simulation tasks show that the proposed self-guidance enables near-optimal performance at negligible inference cost. Notably, under a tight sampling budget, our method achieves up to 70% higher success rates than existing counterparts on challenging dynamic tasks.

Submission Number: 34