Article

Study on confidence calibration for VLA models proposes two lightweight correction techniques

This study presents the first systematic study of confidence calibration for vision-language-action (VLA) foundation models, establishes a baseline for VLA confidence estimation, examines the relationship between task success rate and calibration error as well as how calibration changes over time, and proposes two lightweight techniques, prompt ensembles and action-wise Platt scaling, to correct the observed miscalibration. The paper is 34 pages long with 7 tables and 22 figures, and has been published in TMLR as a final version.

Part of the event: Research on Confidence Calibration of VLA Models Published in TMLR

This page provides a summary and source link only.