Hide-and-Seek in Trajectories: Discovering Failure Signals for VLA Runtime Monitoring
This work is significant for improving the reliability and robust deployment of embodied robotic systems by enabling real-time failure detection for VLA models.
This paper addresses the problem of detecting execution failures in Vision-Language-Action (VLA) models for robots. The proposed Hide-and-Seek framework, using only trajectory-level supervision, localizes failure-indicative actions and achieves state-of-the-art multi-task failure detection performance with a practical accuracy-timeliness trade-off.
Vision-Language-Action (VLA) models enable robots to follow natural language instructions and generalize across diverse tasks, but they remain vulnerable to execution failures that compromise reliability in real-world deployment. Detecting such failures during execution is therefore critical for the robust deployment of embodied systems. Existing failure detection methods either rely on expensive action resampling or external models, while alternatives propagate trajectory-level labels uniformly across every timestep, obscuring localized failure signals. In this paper, we propose \textbf{Hide-and-Seek}, a framework that formulates VLA failure detection as a coarsely supervised learning problem. By combining inter-trajectory and intra-trajectory contrastive objectives, Hide-and-Seek localizes failure-indicative actions and induces temporally structured failure signals from trajectory-level supervision alone, without any step-level annotation. We evaluate Hide-and-Seek on LIBERO, VLABench, and a real-world robotic platform across three representative VLA policies: OpenVLA, $π_0$, and $π_{0.5}$.Our method achieves state-of-the-art multi-task failure detection performance with a practical accuracy--timeliness trade-off under conformal prediction, and generalizes well to both seen and unseen tasks.