Combining Self-Supervised Learning and Imitation for Vision-Based Rope Manipulation | Read Paper on Bytez