11. In Assignment 3 you considered the swish and softplus functions, differentiable approxi- mations to the rectifier function used in neural networks. An approximation that improves on these functions is the mish function 2 log(1+ez) 1 f(x) = x 2 log(1+ez) + 1 (a) [2 marks] Find the intercepts of f(x). (b) [3 marks] What polynomial best approximates f(x) for large negative values of I? (c) [3 marks] What polynomial best approximates f(x) for large positive values of a? (d) [3 marks] The degree 4 Taylor polynomial of f(x) at r = 0 is 00 2 .3 86 C 1 09 25 125 1875 What is f"(0)