Publication
ActionFlow: Efficient, Accurate, and Fast Policies with Spatially Symmetric Flow Matching
N. Funk; J. Urain; J. Carvalho; V. Prasad; G. Chalvatzaki; Jan Peters
In: R:SS workshop: Structural Priors as Inductive Biases for Learning Robot Dynamics. Robotics: Science and Systems (RSS-2024), R:SS, 2024.
Abstract
Most robotic tasks require a proper understanding of the scene’s 3D geometry. Despite the impressive results of deep generative models in complex manipulation tasks, the lack of a representation that encodes the intricate spatial relations between observations and actions usually leads to small spatial generalization capabilities, requiring large amounts of demonstrations. To tackle this problem, we introduce a novel policy class, ActionFlow. ActionFlow integrates spatial symmetry inductive biases while generating expressive action sequences. On the representation level, ActionFlow introduces an SE(3) Invariant Transformer architecture, which enables informed spatial reasoning based
on the relative SE(3) poses between observations and actions. For action generation, ActionFlow leverages Flow Matching, a novel, state-of-the-art deep generative model that has been shown to generate high-quality samples with fast inference, an essential property for feedback control. In combination, ActionFlow policies exhibit strong spatial and locality biases and SE(3) equivariant action generation. Our experiments demonstrate the effectiveness of ActionFlow and its two main components on several simulated and real-world robotic manipulation tasks and confirm that we can obtain efficient, accurate, and fast policies with spatially symmetric flow matching. Project website:
https://flowbasedpolicies.github.io/