This is a PyTorch implementation of the ICLR-20 paper:
By Mitchell Wortsman, Kiana Ehsani, Mohammad Rastegari, Ali Farhadi and Roozbeh Mottaghi (Oral Presentation at CVPR 2019).
asdasdasd Visual Representation for Semantic Target Driven Navigation
As an attempt towards assessing the robustness of embodied navigation agents, we propose RobustNav , a framework to quantify the performance of embodied navigation agents when exposed to a wide variety of visual – affecting RGB inputs – and dynamics – affecting transition dynamics – corruptions.…