DualCoT-VLA: Visual-Linguistic Chain of Thought via Parallel Reasoning for Vision-Language-Action Models
2026-03-23 • Computer Vision and Pattern Recognition
Computer Vision and Pattern RecognitionRobotics
AI summaryⓘ
The authors studied how robots can better understand and follow complex instructions by combining visual details and logical planning. They created a new method called DualCoT-VLA that uses two types of thinking steps at the same time: one focused on detailed visual understanding and another on high-level task planning. This approach also speeds up the robot’s decision-making by doing reasoning steps in parallel instead of one by one. Their experiments showed that this method works well on challenging robot task benchmarks and even in real robot scenarios.
Vision-Language-Action (VLA) modelsChain-of-Thought (CoT) reasoningmulti-modal reasoningspatial perceptionlogical planningautoregressive decodingparallel reasoningrobotic manipulationrobotic benchmarks
Authors
Zhide Zhong, Junfeng Li, Junjie He, Haodong Yan, Xin Gong, Guanyi Zhao, Yingjie Cai, Jiantao Gao, Xu Yan, Bingbing Liu, Yingcong Chen, Liuqing Yang, Haoang Li
Abstract
Vision-Language-Action (VLA) models map visual observations and language instructions directly to robotic actions. While effective for simple tasks, standard VLA models often struggle with complex, multi-step tasks requiring logical planning, as well as precise manipulations demanding fine-grained spatial perception. Recent efforts have incorporated Chain-of-Thought (CoT) reasoning to endow VLA models with a ``thinking before acting'' capability. However, current CoT-based VLA models face two critical limitations: 1) an inability to simultaneously capture low-level visual details and high-level logical planning due to their reliance on isolated, single-modal CoT; 2) high inference latency with compounding errors caused by step-by-step autoregressive decoding. To address these limitations, we propose DualCoT-VLA, a visual-linguistic CoT method for VLA models with a parallel reasoning mechanism. To achieve comprehensive multi-modal reasoning, our method integrates a visual CoT for low-level spatial understanding and a linguistic CoT for high-level task planning. Furthermore, to overcome the latency bottleneck, we introduce a parallel CoT mechanism that incorporates two sets of learnable query tokens, shifting autoregressive reasoning to single-step forward reasoning. Extensive experiments demonstrate that our DualCoT-VLA achieves state-of-the-art performance on the LIBERO and RoboCasa GR1 benchmarks, as well as in real-world platforms.