The Unreasonable Effectiveness of Discrete-Time Gaussian Process Mixtures for Robot Policy Learning

TL;DR

MiDiGap is a novel Gaussian process mixture approach enabling efficient, flexible robot policy learning from few demonstrations, with strong generalization and state-of-the-art results in manipulation tasks.

Contribution

Introduces MiDiGap, a new Gaussian process mixture model for robot policy learning that generalizes across tasks and enables inference-time steering for obstacle avoidance and transfer.

Findings

Achieves state-of-the-art few-shot manipulation performance.

Significantly improves success rates and efficiency in constrained and multimodal tasks.

Enables effective cross-embodiment policy transfer.

Abstract

We present Mixture of Discrete-time Gaussian Processes (MiDiGap), a novel approach for flexible policy representation and imitation learning in robot manipulation. MiDiGap enables learning from as few as five demonstrations using only camera observations and generalizes across a wide range of challenging tasks. It excels at long-horizon behaviors such as making coffee, highly constrained motions such as opening doors, dynamic actions such as scooping with a spatula, and multimodal tasks such as hanging a mug. MiDiGap learns these tasks on a CPU in less than a minute and scales linearly to large datasets. We also develop a rich suite of tools for inference-time steering using evidence such as collision signals and robot kinematic constraints. This steering enables novel generalization capabilities, including obstacle avoidance and cross-embodiment policy transfer. MiDiGap achieves state-of-the-art performance on diverse few-shot manipulation benchmarks. On constrained RLBench tasks, it improves policy success by 76 percentage points and reduces trajectory cost by 67%. On multimodal tasks, it improves policy success by 48 percentage points and increases sample efficiency by a factor of 20. In cross-embodiment transfer, it more than doubles policy success. We make the code publicly available at https://midigap.cs.uni-freiburg.de.