Start Right, Arrive Right: Asynchronous Execution via Initial Noise Selection
For robot learning practitioners, PAINT offers a training-free, gradient-free method to fix chunk boundary inconsistencies in asynchronous execution, improving real-time control without policy modification.
PAINT solves the latency-induced inconsistency problem in asynchronous action chunking for robot policies by selecting an appropriate initial noise before generation, achieving coherent next chunks without modifying the flow ODE. It improves execution consistency and task performance across 12 simulated benchmarks and 6 real-world manipulation tasks.
Action chunking enables robot policies to produce temporally coherent behavior, but generating multi-step action sequences with flow-based policies incurs latency that is incompatible with real-time control. Under asynchronous execution, the robot continues executing the current chunk while the next one is generated, causing even minor delays to create inconsistencies at chunk boundaries. Existing methods address this problem by steering generation toward the already executed action prefix. We instead show that prefix consistency can be achieved by selecting an appropriate initial noise before generation begins, allowing the unmodified flow ODE to produce a coherent next chunk. This reframes asynchronous inference as a noise selection problem rather than a trajectory steering problem. We introduce \textbf{PAINT}, a training-free method that finds this noise via backward Euler inversion and constructs the final chunk through a repainting rule. In summary, \texttt{PAINT} requires no gradients, retraining, or policy modification; yet it improves execution consistency and task performance across \textit{12 simulated benchmarks} and \textit{6 real-world manipulation tasks} spanning single-arm, bimanual, and humanoid embodiments. Website: ~\href{https://paint-action-chunking.github.io}{\texttt{https://paint-action-chunking.github.io}}.