Real-Time Robot Execution with Masked Action Chunking

2026-01-27

AI Summary

This work addresses the issue of robotic execution failures caused by intra-block inconsistency in asynchronous inference. To enhance both internal coherence and temporal continuity of action sequences without introducing additional latency, the authors propose a masked action chunking strategy combined with a prefix-preserving sampling mechanism. This approach explicitly models and resolves intra-chunk inconsistency for the first time, and when integrated with policy fine-tuning, significantly improves task success rates and execution speed in both simulation and real-world environments. The method demonstrates robust performance across varying inference delays, effectively balancing real-time responsiveness with execution reliability.

Technology Category

• Reasoning under Uncertainty: Sequential Decision Making
• Humans and AI: Human-Aware Planning and Behavior Prediction
• Intelligent Robots: Learning & Optimization for ROB

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Search and Retrieval-Augmented AI: Agentic search
• Economics, Online Markets and Human Computation: LLM based quality controls for crowd work

Abstract

Real-time execution is essential for cyber-physical systems such as robots. These systems operate in dynamic real-world environments where even small delays can undermine responsiveness and compromise performance. Asynchronous inference has recently emerged as a system-level paradigm for real-time robot manipulation, enabling the next action chunk to be predicted while the current one is being executed. While this approach achieves real-time responsiveness, naive integration often results in execution failure. Previous methods attributed this failure to inter-chunk discontinuity and developed test-time algorithms to smooth chunk boundaries. In contrast, we identify another critical yet overlooked factor: intra-chunk inconsistency, where the robot's executed action chunk partially misaligns with its current perception. To address this, we propose REMAC, which learns corrective adjustments on the pretrained policy through masked action chunking, enabling the policy to remain resilient under mismatches between intended actions and actual execution during asynchronous inference. In addition, we introduce a prefix-preserved sampling procedure to reinforce inter-chunk continuity. Overall, our method delivers more reliable policies without incurring additional latency. Extensive experiments in both simulation and real-world settings demonstrate that our method enables faster task execution, maintains robustness across varying delays, and consistently achieves higher completion rates.

Problem

• real-time execution
• asynchronous inference
• intra-chunk inconsistency
• robot manipulation
• action chunking

Innovation

• masked action chunking
• asynchronous inference
• intra-chunk inconsistency
• real-time robot execution
• prefix-preserved sampling