STAIRCASE POLICY is introduced, a streaming inference and training framework that turns a flow-matching VLA into a JEPA-style WAM and partitions a large action chunk into sub-chunks at staggered denoising stages, enabling long-horizon execution without repeated full policy inference.
Abstract
World-Action Models (WAMs) improve robotic manipulation by conditioning action generation on predicted future observations, but future prediction adds further inference overhead to already expensive iterative action generation. Action chunking can amortize this cost over multiple actions, yet performance degrades over long execution horizons because later actions remain conditioned on stale observations. We introduce STAIRCASE POLICY, a streaming inference and training framework that turns a flow-matching VLA into a JEPA-style WAM and partitions a large action chunk into sub-chunks at staggered denoising stages. Near-term actions are executed as soon as they become available, while later actions continue to be refined. At each sub-chunk boundary, the future latent is re-predicted from the latest observation and used to update all unexecuted actions, enabling long-horizon execution without repeated full policy inference. The resulting future-prediction error can further serve as a signal for adaptive chunking. S-WAM achieves 97.7% on LIBERO and 87.9% on LIBERO-Plus, and improves performance across multiple policy backbones and real-robot tasks. It reaches 292.7 executed actions per second, $3.62\times$ the throughput of conventional execution at comparable accuracy, while reducing time-to-first-action from 123.6 to 73.3 ms. With additional inference optimizations, throughput further increases to 642.9 actions per second.
Action-chunking vision-language-action (VLA) models predict sequences of future robot actions. Under fixed-horizon execution, the robot executes a fixedlength prefix of each predicted chunk before invoking the policy again, leaving the remaining predictions unused. For a given number of executed actions, short executio...
Jehun Lee, Ha-Jun Lee, Jimin Ryu et al.· 0 citations
World Action Models (WAMs), often built upon pretrained video-generation world models, have emerged as a promising framework for general-purpose and open-ended robot control. However, their real-world deployment often suffers from discontinuous and jerky execution. Existing methods typically address this issue by eithe...
Shao-Jie Ren, Jing-Wen Sun, Ze-Zhi Liu et al.· 0 citations
Action chunking is a standard execution strategy in modern Vision-Language-Action (VLA) frameworks, but fixed execution horizons impose a trade-off between efficiency and accuracy. Short chunks require frequent inference and may cause oscillatory behavior, whereas long chunks can become misaligned with newly observed s...
Run-Ze Xu, Xiaolong Shan, Shuang Dai et al.· 5 citations
SPACE induces two-level programmatic skills from successful trajectories, where subskill boundaries serve as direct chunk-boundary supervision from trajectory-induced programmatic skills, and distilled into a primitive-chunk policy via hybrid on-/off-policy optimization.
Yan-Ting Yang, Can Jin, Jin-Man Zhao et al.· 3 citations
World action models (WAMs) that use future visual prediction at inference time incur substantial generation costs. Asynchronous execution reduces waiting by overlapping inference with robot motion, but visual predictions used for subsequent action generation must anticipate the effects of actions already scheduled for...
Xu-Yao Huang, Yi-Xuan Wang, Zeng-Yao Ye et al.· 1 citation
Rolling-WAM is presented, a formulation that distributes joint denoising across successive replanning cycles and delivers a 4.5x steady-state replanning speedup over standard joint WAMs.
Ying Zhou, Jun-Jie Ye, Yi-Qi Zhao et al.· arXiv.org· 2 citations
From feet to fingertips — we are teaching robots intelligent whole-body control, fine dexterity, and teamwork to complete a broad range of complex tasks.