Uncertainty guided refinement improves robot action chunking success
Uncertainty-Guided Sparse Refinement for Action Chunking Transformer Policies
Robotics
Summary
Robots often plan actions in chunks to handle complex tasks, but mistakes usually happen at just a few key moments, not evenly across the whole plan. The authors introduced a way to spot which steps are uncertain in a robot’s planned action and only fix those steps, instead of redoing the entire plan. This targeted approach made robot teamwork tasks more successful and efficient. Their tests showed better results compared to previous methods across multiple robot manipulation tasks.
What this means in practice
- •For robotics engineers: Improve robot manipulation systems by refining only uncertain steps in planned action sequences for better task success rates.
- •For automation system integrators: Deploy visual robot controllers that selectively correct uncertain actions to increase reliability in dual-arm coordination tasks.
Authors
Chenyang Wang, Yuntian Wang, Xiaoxiong Yang, Dingde Jiang, Siao Liu, Yang Liu
Abstract
Learning chunk-based visuomotor policies for long-horizon robot manipulation remains challenging. Recent action-chunking methods have shown promising performance by predicting temporally extended action sequences. However, their failures are often dominated by prediction errors at a small number of critical timesteps rather than uniformly poor predictions across the entire action chunk, making uniform refinement inefficient and insufficiently targeted. To address this bottleneck, we propose Uncertainty-Guided Refinement (UGR), a sparse refinement framework for chunk-based visuomotor policies. Specifically, UGR follows a coarse-to-refine design: it first predicts a full action chunk, estimates per-step temporal uncertainty from the coarse hidden states, and applies residual correction only to the most uncertain timesteps selected by a binary mask. The uncertainty branch is decoupled from the coarse action predictor, enabling clean attribution of the refinement gains to uncertainty-guided correction rather than additional predictor capacity. Extensive experiments on five dual-arm manipulation tasks from the RoboTwin benchmark show that UGR achieves the best success rate on four tasks, improves over the ACT baseline by up to 13% absolute, and outperforms both full-chunk and position-agnostic block refinement in ablation studies.