SURE: Spatiotemporal Uncertainty Representation for Efficient Vision-Language-Action Learning
SURE:Spatiotemporal Uncertainty Representation for Efficient Vision-Language-Action Learning
This research introduces SURE, a data-efficient training framework for vision-language-action (VLA) policies that explicitly models spatial and temporal uncertainty in action spaces. By augmenting a VLA backbone with a probabilistic action decoder using local correlations and employing Soft-DTW alignment supervision to handle timing variations, SURE reduces sensitivity to demonstrator-specific poses and execution speeds. Evaluations on LIBERO, RoboTwin 2.0, and real-world manipulation tasks show that SURE achieves success rates comparable to or better than baseline VLAs even when using only 10% of demonstration data or 20% of training steps.
The paper proposes SURE, a framework that models spatiotemporal uncertainty in action spaces to improve data efficiency for vision-language-action (VLA) policies. It claims matching or surpassing baseline performance with only 10% of demonstration data or 20% of training steps on LIBERO, RoboTwin 2.0, and real-world tasks.
Source: arXiv Robotics — research abstracts · Read original article ↗
Loading article text…
What the source reports
Publisher-reported claims, with original evidence. These results have not been independently verified by RoboSignal.
- paper: not_reported
Source excerpts and review record
Automatically extracted; no manual editorial approval recorded.
evaluate SURE on LIBERO, RoboTwin 2.0, and real-world manipulation tasks. Even with only 10% of the demonstration data or 20% of the training steps, SURE achieves manipulation success rates that match or surpass baseline VLAs.
Open source S5
Source：arXiv Robotics — research abstracts · arxiv.org