One paper got accepted by CoRL 2024, enabling generating realistic and challenging environments for training generalizable robot navigation policies.