AI RESEARCH
FLASH: Efficient Visuomotor Policy via Sparse Sampling
arXiv CS.CV
•
ArXi:2605.15492v1 Announce Type: cross Generative models such as diffusion and flow matching have become dominant paradigms for visuomotor policy learning, yet their reliance on iterative denoising incurs high inference latency incompatible with real-time robotic control. We present Fast Legendre-polynomial Action policy via Sparse History-anchored flow (FLASH Policy), which replaces discrete action-chunk generation with continuous Legendre polynomial trajectory representation.