Search papers, labs, and topics across Lattice.
This paper introduces Factored Diffusion Policies (FDP), a novel approach to robot control that uses a single diffusion network trained with per-factor null-token dropout to achieve compositional generalization across task factors. By assuming approximate conditional independence between factors, FDP decomposes the score function additively at inference, significantly reducing the training data required compared to training separate networks for each factor combination. Experiments on drone racing demonstrate that FDP achieves strong generalization, matching an oracle policy on multi-gate racing and exhibiting superior zero-shot transfer to unseen environments compared to baseline compositional methods.
Forget training separate networks for every robotic task variation – this method uses a single diffusion network to generalize to new combinations of factors, achieving near-oracle performance in drone racing.
Robotic tasks are typically specified by a tuple of factors, such as the object to be grasped, the obstacles to be avoided, the color of the target, and so on. Collecting expert demonstrations for every combination of factor values grows combinatorially. We present factored diffusion policies: a single shared diffusion network trained with per-factor null-token dropout, whose score decomposes additively across factors at inference. Under approximate conditional independence between factors given the action-observation pair, this composition approximates the true joint score with a bounded uniform error, reducing the training-task budget from a product of factor cardinalities to a sum. A trajectory-tube certificate chains this score-level bound through the reverse-time sampling ODE and a contracting tracking controller into a closed-loop state-trajectory tube whose radius factors into an ODE-sensitivity constant and a per-factor score-error budget. Unlike compositional-diffusion methods for control that combine separately trained networks, we use one shared network. Drone racing experiments confirm both the generalization bound and the certificate. On state-based multi-gate racing, the factored policy passes 90% of held-out gates -- matching an oracle -- while a K-network composition baseline collapses to 3%; on vision-based single-gate traversal, it transfers zero-shot to an unseen venue with +11.7pp success-rate gain and 2.4X crash-rate reduction.
