Time-Equivariant Contrastive Video Representation Learning
                        Simon Jenni  Hailin Jin
                          Adobe Research
                        {jenni,hljin}@adobe.com

             Abstract
  We introduce a novel self-supervised contrastive learn-
ing method to learn representations from unlabelled videos.
Existing approaches ignore the specifics of input distor-
tions, e.g., by learning invariance to temporal transforma-
tions. Instead, we argue that video represent ...