作者位次 #4/5ACL 2026Poster· Natural Language Processing
From Off-Policy to On-Policy: Enhancing GUI Agents via Bi-level Expert-to-Policy Assimilation
Zezhou Wang · Ziyun Zhang · Xiaoyi Zhang · Zhuzhong Qian · Yan Lu
摘要
Vision-language models are increasingly deployed as computer-use agents (CUAs) that operate desktops and browsers. Top-performing CUAs are framework-based systems that decompose planning and execution, while end-to-end screenshot-to-action policies are easier to deploy but lag behind on benchmark...