Refining Compositional Diffusion for Reliable Long-Horizon Planning
For long-horizon planning in robotics, this training-free guidance method improves reliability of compositional diffusion without requiring additional training.
Compositional diffusion planning suffers from mode-averaging when local plan distributions are multimodal, leading to infeasible trajectories. The proposed RCD method uses self-reconstruction error and overlap consistency to guide sampling toward high-density, globally coherent plans, outperforming existing methods on OGBench tasks.
Compositional diffusion planning generates long-horizon trajectories by stitching together overlapping short-horizon segments through score composition. However, when local plan distributions are multimodal, existing compositional methods suffer from mode-averaging, where averaging incompatible local modes leads to plans that are neither locally feasible nor globally coherent. We propose Refining Compositional Diffusion (RCD), a training-free guidance method that steers compositional sampling toward high-density, globally coherent plans. RCD leverages the self-reconstruction error of a pretrained diffusion model as a proxy for the log-density of composed plans, combined with an overlap consistency term that enforces consistency at segment boundaries. We show that the combined guidance concentrates sampling on high-density plans that mitigate mode-averaging. Experiments on challenging long-horizon tasks from OGBench, including locomotion, object manipulation, and pixel-based observations, demonstrate that RCD consistently outperforms existing methods.