arXiv:2611.09684v1 Announce Type: cross
Abstract: Vision-language-action (VLA) policies have become a promising paradigm for end-to-end robot control, but their effectiveness still depends heavily on large volumes of demonstration data. In manipulation tasks, this dependence is compounded by structured spatiotemporal uncertainty in the action space: multiple nearby actions or trajectories may successfully solve the same task, while semantically equivalent demonstrations often unfold at different speeds. Standard pointwise supervision treats each demonstrated action as an exact target, imposing an unnecessary learning burden under such uncertainty. We present SURE, a data-efficient VLA training framework that explicitly models both spatial and temporal uncertainty. SURE augments a VLA backbone with a lightweight transposed-convolutional probabilistic action decoder that captures spatial uncertainty through local correlations among neighboring action bins. It further introduces Soft-DTW alignment supervision to accommodate local timing variations between predicted and demonstrated action chunks. Together, these two components reduce sensitivity to demonstrator-specific poses and execution timing, enabling more data-efficient action learning. We 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.
arxiv
SURE: Spatiotemporal Uncertainty Representation for Efficient Vision-Language-Action Learning
SURE is a data-efficient training framework for vision-language-action policies that explicitly models spatial and temporal uncertainty in manipulation demonstrations. It adds a lightweight transposed-convolutional probabilistic action decoder to capture spatial uncertainty among neighboring action bins and uses Soft-DTW alignment supervision to accommodate local timing variations. The method is evaluated on LIBERO, RoboTwin 2.0, and real-world manipulation tasks, where it matches or surpasses baseline VLA success rates using only 10% of demonstration data or 20% of training steps.
It presents a concrete training method for reducing the demonstration data and training steps required by vision-language-action robot policies.
Comments & Questions (0)
No comments yet
Be the first to comment!