1 paper · 1 filter
Yuhao Cao, Xiao Liu, Yang Xie +2
Most existing vision-language navigation tasks assume that instructions are complete and unambiguous. However, real-world robots often encounter natural human instructions that are…