Self-Supervised Multi-View Learning via Auto-Encoding 3D Transformations | Read Paper on Bytez