A Constructive Approach for One-Shot Training of Neural Networks Using Hypercube-Based Topological Coverings | Read Paper on Bytez