Consider the following training dataset: We would like to fit a linear regression model y^=w0+w1x to this dataset using gradient lescent and 2 regularization, with =0.1 and =0.1. We use mean squared error for our oss function and initialize w=(w0,w1)=0 Part(a): Perform the first two gradient updates to w (i.e. what are the weights after one gradient update? after two?). Part(b) : Perform the first two gradient updates to w with momentum using =0.9 (assume z(0)=0 for the first gradient update). Part(c) : What would happen if we set to be much smaller, e.g. =10100 ? What about much larger, e.g. =100 ? (You do not need to perform the computations manually; a written description is sufficient.)