KERM: Knowledge Enhanced Reasoning for Vision-and-Language Navigation | Read Paper on Bytez