作者位次 #5/11ACL 2026Poster· Natural Language Processing
Don’t Act Blindly: Robust GUI Automation via Action-Effect Verification and Self-Correction
Yuzhe Zhang · Xianwei Xue · Xingyong Wu · Mengke Chen · Chen Liu · Xinran He · Run Shao · Feiran Liu · Huanmin Xu · Qiutong Pan · Haiwei Wang
摘要
Autonomous GUI agents based on vision-language models (VLMs) often assume deterministic environment responses, generating actions without verifying whether previous operations succeeded. In real-world settings with network latency, rendering delays, and system interruptions, this assumption leads...