CSE 573 - Introduction to Artificial Intelligence

University of Washington Department of Computer Science & Engineering

CSE 573 - Introduction to Artificial Intelligence - Autumn 2012
Wed, Fri 1:30-2:50pm in EEB 025 (moved from 042)

CSE Home

About Us

Contact Info

Instructor: Dan Weld (weld at cs dot washington dot edu)
Office hours: Thurs 9am in CSE 588 or by email TA: Johnson Goh (johnson at cs dot washington dot edu)
Office hours: Wed 10:30am in CSE 002 computer cluster

Schedule

Date	Topics & Lecture Notes	Readings
September 26	Introduction, Agents, Problem Spaces	R&N Ch. 2; (Optional Ch 1)
September 28	Heuristic Search	R&N Ch3 thru 3.5
October 3	Generating Heuristics & Intro to Games	R&N Sections 3.6 & 5.1
October 5	Mini-max, alpha-beta & expectimax	R&N Finish Ch 5 & read Ch 13
October 10	decision theory & MDPs	R&N Sections 16.1-16.4, 16.6, 17.1-17.4
October 12	MDPs: value iteration, RTDP & policy iteration
October 17	Monte Carlo Planning: Multi-armed bandit, policy rollout, sparse sampling, adaptive MCP, UCB, UCT	Sections 3.1-3.3 of Kloetzer's thesis
October 19	Reinforcement Learning: ADP, TD-learning, Q-learning, credit assignment, exploration / exploitation policies	R&N Ch 21 thru 21.3; Optional Szepesvari tutorial
October 26	Reinforcement Learning II - approximation techniques	R&N Sections 21.4 - 21.7
October 31	Introduction to uncertainty & HMMs	R&N Chapter 13 & Sections 14.1, 14.2, 15.1 & 15.2
November 2	Particle filters for HMMs	R&N Sections 14.5 & 15.5
November 7	Bayesian networks: semantics, independence & inference
November 9	Learning Bayesian networks	R&N Sections 18.1, 18.2, 20.1, 20.2
November 14	Structure Learning, EM, Ensembles	R&N Sections 18.10, 18.11, 20.3
November 28	POMDPs and their applications	R&N Section 17.4
November 30	Constraint Satisfaction Problems	R&N Chapter 6
December 5	Knowledge Representation: Propositional Logic, First-order logic & MLNs	R&N Chapters 7,8 & 9
December 7	Project Presentations
December 14	Final reports due by 1:30pm

Programming Projects

This quarter, we will use the excellent Berkeley Pac-Man Projects originaly developed by John DeNero and Dan Klein. Please complete the versions listed below, as they differ in places from the originals. Use Python 2.x

Project 0: Python Tutorial (Not Graded)
Project 1: Search in Pac-man (Due Thursday 11 October 2012 11:45pm in Dropbox) - Corrected version may be turned in for 50% of missed points by Monday 10/22 (ie your final grade will be the average of your original score and the score received for a corrected version
Project 2: MultiAgent Pac-man (Due Thursday 18 October 2012 11:45pm in Dropbox) - Corrected version may be turned in for 50% of missed points by Monday 11/5
Project 3: MDPs and RL (Due Tues 30 October, 11:45pm) - Corrected version may be turned in for 50% of missed points by Monday 11/5
Project 4: Ghostbusters (Inference and Filtering) (Due Monday Nov 12, 11:45pm)

Project

Check here for info on the final project. Tentative group definitions and a one-sentence idea due Wed 11/14.

Textbooks

Required: Stuart Russell & Peter Norvig, Artificial Intelligence: A Modern Approach, Prentice-Hall, Third Edition (2009) [R&N].
Useful: Richard Sutton & Andrew Barto, Reinforcement Learning: An Introduction, MIT Press. (limited chapters; freely available online) [S&B]

Course Administration and Policies

Your grade will be 40% programming assignments, 10% written homeworks, 10% class participation and 30% final project. No midterm or final.
Assignments should be done individually unless otherwise specified. You may discuss the subject matter with other students in the class, but all final answers must be your own work. You are expected to maintain the utmost level of academic integrity in the course.
Assignments may be handed in up to five days late, at a penalty of 10% of the maximum grade per day.

Communication

Please feel free to email Dan or Johnson anytime or stop by their offices whenever convenient. Note: Dan is often in a panic, preparing with door shut immediately before class, but other times are great - or email for an appointment.
If you have comments of general interest, please use the discussion forum
For feedback to course staff, contact us directly or send an anonymous email.
Deposit homework in the dropbox

Department of Computer Science & Engineering
University of Washington
Box 352350
Seattle, WA 98195-2350
(206) 543-1695 voice, (206) 543-2969 FAX