研究人员开发了GUICrafter，一种新颖的弱监督GUI智能体，旨在克服训练此类系统中的数据收集挑战。通过利用海量未标注截图，GUICrafter显著减少了对昂贵人工标注的需求。该智能体采用两阶段课程学习框架，首先从未标注数据中学习视觉基础，然后通过强化学习用少量高质量数据进行校准。实验表明，GUICrafter使用的数据量仅为现有系统（如UI-TARS）的一小部分，即可达到与之相当或更好的性能。 AI
影响 这种方法通过降低数据标注成本，可能显著降低开发复杂GUI智能体的入门门槛。
排序理由 该集群描述了一篇详细介绍新型AI模型及其方法论的研究论文。
AI 生成摘要 · Google Gemini · 来自 3 个来源。 我们如何撰写摘要 →
arXiv:2606.29705v1 Announce Type: new Abstract: Data, as the fundamental substrate of modern intelligence, has greatly driven the development of current foundation models. Naturally, researchers aim to extend this paradigm to the domain of GUI agents, hoping to build strong GUI a…
Data, as the fundamental substrate of modern intelligence, has greatly driven the development of current foundation models. Naturally, researchers aim to extend this paradigm to the domain of GUI agents, hoping to build strong GUI agents through a similar paradigm. However, GUI a…
GUICrafter addresses GUI agent data challenges through a weakly-supervised approach using unannotated screenshots and a two-stage curriculum learning framework for visual grounding and reinforcement learning calibration.