\ CSE 473 - Introduction to Artificial Intelligence - Autumn 2014
Steam-powered Turing Machine University of Washington Department of Computer Science & Engineering
CSE 473 - Introduction to Artificial Intelligence - Autumn 2014
Mon, Wed, Fri 9:30-10:20 in Hitchcock 132
  CSE Home  About Us    Search    Contact Info 

Instructor: Dan Weld
Office hours: Fri 10:30 in CSE 588 or by email 		TA: Galen Andrew
Office hours: Wed 1:00-3:00 in CSE 218
TA: Naozumi Hiranuma
Office hours: Tue 1:30-2:30 in CSE 218, Thu 1:00-2:00 in CSE 220
TA: Travis Mandel
Office hours: Fri 3:30-4:30 in CSE 218
TA: Jeff Shaffer
Office hours: Wed 10:30-11:30 in CSE 021

This term we will be using Piazza for class discussion. The system is highly catered to getting you help fast and efficiently from classmates, the TA, and myself. Rather than emailing questions to the teaching staff, I encourage you to post your questions on Piazza. If you have any problems or feedback for the developers, email team@piazza.com.

Find our class page at: https://piazza.com/washington/fall2014/cse473/

First time logging in? You should have gotten an activation email from the Piazza system. Alternatively, visit the link above and follow the instructions to enroll as a student in CSE 473.

Schedule

Date Topics & Lecture Notes Readings
September 24 Introduction, Agents Optional: R&N, Ch. 1 & Ch. 2
September 26 Problem Spaces & Blind Search R&N, Ch. 3 thru p91
September 29 Best-First, Uniform-Cost, Greedy & A* Search R&N, Ch. 3, p 92-98
October 1 Heuristics & Pattern DBs R&N, Section 3.6, (optional: pattern DB paper)
October 3 Local Search R&N, Section 4.1
October 6 Constraint Satisfaction Problems I (.pdf, .pptx with demos) R&N, Chapter 6
October 8 CSPs, part II None
October 10 Adversary search, min-max R&N Chapter 5 thru p174
October 13 Alpha-beta search R&N, 5.3 & 5.4
October 15 Expectimax search R&N, 5.5, 5.7 & 5.9
October 17 Markov decision processes (MDPs) R&N, section 17.1
October 20 Bellman equations (see previous slides)
October 22 MDPs: value iteration R&N, section 17.2
October 24 MDPs: policy iteration (ppt) R&N, section 17.2
October 27 Midterm (Solutions) Practice: Probs 1, 2, 3, 8b, 8c Probs 1a, 2, 3 R&N problems 6.4, 6.5
October 29 Reinforcement Learning R&N, section 21.1, 21.2
October 31 RL Continued (Q-Learning) R&N, section 21.1, 21.2
November 3 RL Continued (Approximate Q-Learning)/Uncertainty R&N sections 21.3 & 21.4
November 5 Uncertainty: Inference by enumeration & Bayes Rule R&N Ch 13
November 7 Markov Models R&N Sections 15.1 & 15.2
November 10 Hidden Markov Models R&N Sections 15.3
November 12 No class
November 14 Hidden Markov Models & the Forward Algorithm
November 17 Particle Filters for HMMs R&N 15.5
November 19 Bayes Nets R&N 14.1 & 14.2
November 21 Independence in Bayes Nets R&N 14.3
November 24 Inference in Bayes Nets R&N 14.4
November 26 Variable Elimination, NP Completeness, Polytrees
December 1 Parameter Learning R&N Sections 18.1, 18.2, 20.1, 20.2
December 3 Expectation maximization & learning Bayesnet structure R&N Sections 20.2 & 20.3
December 5 Summary

Course Administration

Academic Dishonesty Policy

It is encouraged that you discuss your solutions with each other and consult online sources to better understand the material. However your code must be written entirely by yourself. As a rule, you should never look at or run anyone else's code for the assignment, whether the code was written by someone currently in the class, or someone who took it previously, even at another university. Reading pseudocode for generic algorithms (like alpha-beta pruning or A* search) is perfectly OK. If you use a source very closely, for example, converting a pseudocode implementation of A* to python, academic integrity demands that you cite the source (in a comment). You will not be penalized for this; on the contrary, the citation may help us to understand why your implementation is so similar to someone else's, in case they use and cite the same source. We do compare everyone's projects to each other and to past submissions to detect logical redundancy. When two assignments are too similar to have occurred by chance, we have to look into whether something improper occurred. These investigations are not fun for anyone involved, so please, be careful to come up with your solutions entirely independently.

Textbooks

Programming Projects

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

Final Exam

Here are some handy study problems
  • The exam itself is Wed 12/10 in class 8:30am -10:20am. Closed book. No calculators, computers or phones. No internet. But you can bring in one 8.5 x 11 inch piece of paper (double sided ok) with anything written or typed on it.

    Communication

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