Search papers, labs, and topics across Lattice.
This paper addresses the challenge of scaling diffusion models to long-horizon robot planning tasks by introducing a compositional approach that enforces boundary agreement on Tweedie estimates (clean data) rather than noisy intermediate states. They formulate long-horizon planning as inference over a factor graph of overlapping video chunks, using pretrained short-horizon video diffusion models as local priors. By employing synchronous and asynchronous message passing on Tweedie estimates, the method achieves globally consistent plans without additional training.
Ditch the noisy averaging: this method unlocks stable, long-horizon robot planning with diffusion models by enforcing consistency on *clean* data estimates.
Diffusion models excel at short-horizon robot planning, yet scaling them to long-horizon tasks remains challenging due to computational constraints and limited training data. Existing compositional approaches stitch together short segments by separately denoising each component and averaging overlapping regions. However, this suffers from instability as the factorization assumption breaks down in noisy data space, leading to inconsistent global plans. We propose that the key to stable compositional generation lies in enforcing boundary agreement on the estimated clean data (Tweedie estimates) rather than on noisy intermediate states. Our method formulates long-horizon planning as inference over a chain-structured factor graph of overlapping video chunks, where pretrained short-horizon video diffusion models provide local priors. At inference time, we enforce boundary agreement through a novel combination of synchronous and asynchronous message passing that operates on Tweedie estimates, producing globally consistent guidance without requiring additional training. Our training-free framework demonstrates significant improvements over existing baselines, effectively generalizing to unseen start-goal combinations that were not present in the original training data. Project website: https://comp-visual-planning.github.io/
