VLASH: Real-Time VLAs via Future-State-Aware Asynchronous Inference

TL;DR

VLASH introduces a future-state-aware asynchronous inference framework for vision-language-action models, enabling faster, smoother, and more accurate robotic control without additional overhead, effectively handling high-speed tasks.

Contribution

VLASH is the first to estimate future robot states during asynchronous inference, bridging the prediction-execution gap and improving reaction speed and stability in VLAs.

Findings

Achieves up to 2.03x speedup over synchronous inference.

Reduces reaction latency by up to 17.4x.

Enables high-precision fast-reaction tasks like ping-pong.

Abstract

Vision-Language-Action models (VLAs) are becoming increasingly capable across diverse robotic tasks. However, their real-world deployment remains slow and inefficient: demonstration videos are often sped up by 5-10x to appear smooth, with noticeable action stalls and delayed reactions to environmental changes. Asynchronous inference offers a promising solution to achieve continuous and low-latency control by enabling robots to execute actions and perform inference simultaneously. However, because the robot and environment continue to evolve during inference, a temporal misalignment arises between the prediction and execution intervals. This leads to significant action instability, while existing methods either degrade accuracy or introduce runtime overhead to mitigate it. We propose VLASH, a general asynchronous inference framework for VLAs that delivers smooth, accurate, and fast reaction control without additional overhead or architectural changes. VLASH estimates the future execution-time state by rolling the robot state forward with the previously generated action chunk, thereby bridging the gap between prediction and execution. Experiments show that VLASH achieves up to 2.03x speedup and reduces reaction latency by up to 17.4x compared to synchronous inference while fully preserving the original accuracy. Moreover, it empowers VLAs to handle fast-reaction, high-precision tasks such as playing ping-pong and playing whack-a-mole, where traditional synchronous inference fails. Code is available at https://github.com/mit-han-lab/vlash