CODA：協調大腦與小腦的雙腦計算機 使用解耦強化學習的代理系統

CODA: Coordinating the Cerebrum and Cerebellum for a Dual-Brain Computer Use Agent with Decoupled Reinforcement Learning

August 27, 2025

作者: Zeyi Sun, Yuhang Cao, Jianze Liang, Qiushi Sun, Ziyu Liu, Zhixiong Zhang, Yuhang Zang, Xiaoyi Dong, Kai Chen, Dahua Lin, Jiaqi Wang

cs.AI

摘要

面向图形用户界面（GUI）的自主代理在科学计算等专业领域面临重大挑战，这些领域既需要长远的规划，又要求精确的执行。现有方法存在一种权衡：通才型代理擅长规划但在执行上表现欠佳，而专才型代理则表现出相反的弱点。近期的组合框架试图通过结合规划者和执行者来弥合这一差距，但这些框架通常是静态且不可训练的，从而阻碍了从经验中适应。鉴于科学领域高质量数据的稀缺性，这是一个关键的限制。为解决这些限制，我们引入了CODA，一种新颖且可训练的组合框架，它将通才型规划者（Cerebrum）与专才型执行者（Cerebellum）集成，通过专门的两阶段流程进行训练。在第一阶段，即专业化阶段，我们采用解耦的GRPO方法，为每个科学应用单独训练一个专家规划者，从一小部分任务轨迹中引导。在第二阶段，即泛化阶段，我们汇集所有来自专业专家的成功轨迹，构建一个整合的数据集，随后用于对最终规划者进行监督微调。这使得CODA既具备稳健的执行能力，又拥有跨领域的泛化能力。在ScienceBoard基准测试的四个挑战性应用上评估，CODA显著超越了基线模型，并在开源模型中确立了新的技术前沿。

English

Autonomous agents for Graphical User Interfaces (GUIs) face significant challenges in specialized domains such as scientific computing, where both long-horizon planning and precise execution are required. Existing approaches suffer from a trade-off: generalist agents excel at planning but perform poorly in execution, while specialized agents demonstrate the opposite weakness. Recent compositional frameworks attempt to bridge this gap by combining a planner and an actor, but they are typically static and non-trainable, which prevents adaptation from experience. This is a critical limitation given the scarcity of high-quality data in scientific domains. To address these limitations, we introduce CODA, a novel and trainable compositional framework that integrates a generalist planner (Cerebrum) with a specialist executor (Cerebellum), trained via a dedicated two-stage pipeline. In the first stage, Specialization, we apply a decoupled GRPO approach to train an expert planner for each scientific application individually, bootstrapping from a small set of task trajectories. In the second stage, Generalization, we aggregate all successful trajectories from the specialized experts to build a consolidated dataset, which is then used for supervised fine-tuning of the final planner. This equips CODA with both robust execution and cross-domain generalization. Evaluated on four challenging applications from the ScienceBoard benchmark, CODA significantly outperforms baselines and establishes a new state of the art among open-source models.