CSE 473 Introduction to Artificial Intelligence
Autumn 2001
Problem Set 5
Due November 26th, 1:30pm.
Reading: Read Chapter 17 on machine learning.
- (6 points) Problem 2.3 on the attached page (see below). Note: the hypothesis space is the one discussed in class. Table 2.1 corresponds to the Table in Slide 2.
- (14 points) Problem 2.4 on the attached page.
- (10 points, extra credit) If necessary, modify your Othello program’s evaluation function to be a polynomial of features and weights of the sort shown in class. Figure out a “training regime” whereby your program can tune the weights on its evaluation function by playing against itself. Please turn in the following:
- Your original feature weights along side with your feature weights after learning. How many features weights changed? By how much?
- A brief description of your “training regime” (how did you train it, how many games did it play? How did you decide to stop self play?)
- An evaluation of improvement due to learning: is your program better now? How do you know? Can you quantify how much better it is?
