Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Temporal Difference Calibration in Sequential Tasks: Application to Vision-Language-Action Models

About

Recent advances in vision-language-action (VLA) models for robotics have highlighted the importance of reliable uncertainty quantification in sequential tasks. However, assessing and improving calibration in such settings remains mostly unexplored, especially when only partial trajectories are observed. In this work, we formulate sequential calibration for episodic tasks, where task-success confidence is produced along an episode, while success is determined at the end of it. We introduce a sequential extension of the Brier score and show that, for binary outcomes, its risk minimizer coincides with the VLA policy's value function. This connection bridges uncertainty calibration and reinforcement learning, enabling the use of temporal-difference (TD) value estimation as a principled calibration mechanism over time. We empirically show that TD calibration improves performance relative to the state-of-the-art on simulated and real-robot data. Interestingly, we show that when calibrated using TD, the VLA's single-step action probabilities can yield competitive uncertainty estimates, in contrast to recent findings that employed different calibration techniques.

Shelly Francis-Meretzki, Mirco Mutti, Yaniv Romano, Aviv Tamar• 2026

Related benchmarks

TaskDatasetResultRank
Failure DetectionLIBERO seen
Brier Score0.061
37
Failure DetectionLIBERO Unseen
Brier Score0.097
37
Failure DetectionWidowX seen
Brier Score0.096
11
Failure DetectionWidowX (unseen)
Brier Score0.153
11
Failure DetectionFranka seen
Brier Score0.15
11
Failure DetectionFranka unseen
Brier Score0.215
11
Showing 6 of 6 rows

Other info

Follow for update