1 paper · 2 filters
Dhruv Sharma, Alihusein Kuwajerwala, Florian Shkurti
The robustness of visual navigation policies trained through imitation often hinges on the augmentation of the training image-action pairs. Traditionally, this has been done by col…