Theorem 1 in the paper is wrong

I understand that you have forced some theoretical results to improve the chances of the paper being accepted, but the derivation and conclusion of Theorem 1 are completely wrong, and I hope you will not mention or cite this theoretical result in the future work. 

I hope you have a more rigorous understanding of the principles of derivative of composite functions and backpropagation.