Self-Supervised Representation Learning for Vision