1. In previous homework (appended at the bottom) you computed the SoftMaX loss L. Now using numerical method compute , the derivative of L with respect to W in matrix form. 2. Next, for a learning rate of 0.2, compute W, Now compute next W.

1. In previous homework (appended at the bottom) you computed the SoftMaX loss L. Now using numerical method compute WL, the derivative of L with respect to W in matrix form. 2. Next, for a learning rate of 0.2 , compute W, Now compute next W. Consider the following linear classification network with 4 pixels and 4 classes \{cat, dog, cow, horse . The weight matrix W is given by: For the following input/labels compute the SVM and SoftMax loss function L.