Dongseok Shim, H. Jin Kim: Learning a Domain-Agnostic Visual Representation for Autonomous Driving via Contrastive Loss. CoRR abs/2103.05902 (2021)