1 Softmax regression and logistic regression [10 points] In lecture 15, we introduced the softmax regression, which learns for each class c a pair (wc,bc) such that p(cx)=cexp(wcx+bc)exp(wcx+bc). Suppose that we only have two classes c=+1 and c=1, please rewrite p(+1x) as 1+exp((wx+b))1 and represent (w,b) in terms of w+1,b+1,w1,b1. That is, a softmax regression reduces to logistic regression when the classification problem is binary. 1 Softmax regression and logistic regression [10 points] In lecture 15, we introduced the softmax regression, which learns for each class c a pair (wc,bc) such that p(cx)=cexp(wcx+bc)exp(wcx+bc). Suppose that we only have two classes c=+1 and c=1, please rewrite p(+1x) as 1+exp((wx+b))1 and represent (w,b) in terms of w+1,b+1,w1,b1. That is, a softmax regression reduces to logistic regression when the classification problem is binary