2. We have mainly focused on squared loss, but there are other interesting losses in data-mining. Consider the following loss function which we denote by 0(2) = max(0, -2). Let S be a training set (2, y),...,x,y) where each r ER" and y E{-1,1}. Consider running stochastic gradient descent (SGD) to find a weight vector w that minimizes 12 oly. wr). Explain the explicit relationship between this algorithm and the Perceptron algorithm. Recall that for SGD, the update rule on the ith example is Wnew = wold - 706 (y'w?:)