1 paper
Weihua Cheng, Junming Liu, Yifei Sun +3
Multimodal Large Language Models (MLLMs) have significantly advanced GUI agents, yet long-horizon automation remains constrained by two critical bottlenecks: context overload from…