Self-supervised Co-training
         for Video Representation Learning

             Tengda Han, Weidi Xie, Andrew Zisserman
         VGG, Department of Engineering Science, University of Oxford
              {htd, weidi, az}@robots.ox.ac.uk
          http://www.robots.ox.ac.uk/~vgg/research/CoCLR/.

                       Abstract
     The objective of this paper is visual-only self-supervised video representation learn-
     ing. We make the following contributions: (i) we investigate the bene ...