Abstract: Most Vision-and-Language Navigation (VLN) algorithms are prone to making inaccurate decisions due to their lack of visual common sense and limited reasoning capabilities. To address this ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results
Feedback