1 paper · 1 filter
LinFeng Li, Jian Zhao, Yuan Xie +2
The dominant paradigm for training Large Vision-Language Models (LVLMs) in navigation relies on imitating expert trajectories. This approach reduces the complex navigation task to…