Mixed Autoencoder for Self-Supervised Visual Representation Learning | Read Paper on Bytez