Breaking Down and Building Up: Mixture of Skill-Based Vision-and-Language Navigation Agents

Tianyi Ma, Yue Zhang, Zehao Wang +1

Vision-and-Language Navigation (VLN) poses significant challenges for agents to interpret natural language instructions and navigate complex 3D environments. While recent progress…