COMPSCI 189 Study Guide - Midterm Guide: Tikhonov Regularization, Gaussian Noise, Grayscale
Document Summary
Give both the binarized input, the true grayscale, and the output of your model. You may use the code from previous part to visualize the images. Figure 2: example results with the input on the left and output on the right. We are given a set of n training samples {xi, yi}n i=1 that are generated by the above model with i. i. d. Our goal is to t a linear model and get an estimate bw for the true parameter w . For all parts, assume that xi"s are given and xed (not random). For a given training set {xi, yi}n i=1, the ridge-regression estimate for w is de ned by bw = arg min w r. W2 + nx i=1 (yi xiw)2 with 0. 7 (a) (8 pts) compute the squared-bias of the ridge estimate bw de ned as follows.
