SUREFlow: State-space Uncertainty-aware REsidual Flow Matching for Robust Robot Manipulation
2026-07-11 • Robotics
Robotics
AI summaryⓘ
The authors present SUREFlow, a new method to help robots perform better at tasks by managing uncertainties in their actions. Unlike older methods that assume equal errors and don't explicitly handle uncertainty, SUREFlow predicts both the robot's next move and the confidence in that move. This allows the robot to focus on improving uncertain parts without needing feedback from the environment, making its actions more reliable over time. Their approach performs much better on robot task benchmarks than previous methods, while using fewer parameters.
robot manipulationgenerative vision-language-action policiesuncertainty modelingresidual flow matchingvelocity errorsstate-space modelsaction refinementLIBERO benchmarkMamba backbone
Authors
Md Tanvir Islam, Sai Navaneet Peddapalli, Sangmoon Lee, Sangtae Ahn
Abstract
Generative vision-language-action policies have advanced robot manipulation, but they often exhibit instability under noise, partial observability, and stochastic initial conditions. During extended rollouts, small velocity errors accumulate, degrading execution reliability. Existing diffusion and flow-based policies typically assume homoscedastic residuals and lack explicit uncertainty modeling within action generation, limiting robustness during iterative rollout. We propose SUREFlow, a state-space uncertainty-aware residual flow matching framework built on a Mamba backbone. The method jointly predicts action velocities and input-dependent residual uncertainty, enabling selective refinement of unreliable action dimensions without environment feedback while preserving computational efficiency. On LIBERO, SUREFlow achieves 92.5% average success rate (SR), outperforming the Mamba-based MaIL by 34.2%. On LIBERO-PRO, it attains around 49% SR using only 179M parameters, achieving performance comparable to large VLAs with 3-7B parameters. SUREFlow source code is available on: https://github.com/tanvirnwu/SUREFlow