CISC 7414x Expert Systems

Last Update: October 16th 2010 1:31pm

Instructor

Yuqing Tang [website]
Email: ytang@cs.gc.cuny.edu
Course website: [http://web.cs.gc.cuny.edu/~tang/teachings/cis7414x]
Office hours: Please email me to schedule an appointment.

Objectives
To give computer science Master’s Degree Candidates a thorough foundation in the discipline of Artificial Intelligence, focusing on Expert Systems, and related methodologies; to bring current trends and advances in the discipline to the forefront so that they may be considered for possible use as solutions to appropriate problems.

In Fall 2010, the course will focus on probabilistic approaches, in particular, Bayesian networks.

Time and Place

Day	Time	Location
W	6:05PM- 8:10PM	236 NE

Textbook

Bayesian Artificial Intelligence (2004), Kevin B. Korb and Ann E. Nicholson, Chapman and Hall, CRC Press [errata]

Supplemental Materials

Internal handouts (please email me for the access password)
Expert Systems: Principles and Programming (4th ed.) Joseph C. Giarratano, Gary D. Riley, Thomson Course Technology
Artificial Intelligence: A Modern Approach (3rd ed.), Stuart Russell, Peter Norvig, Prentice Hall
Bayesian Networks and Decision Graphs (2 ed.), Finn B. Jensen, Thomas Graven-Nielsen, Springer

Class Project

Requirements
Important dates
- October 6th: Class discussion
- October 13rd: Project proposal due
- November 10th: Milestone report due
- December 1st: Final report and presentation due

Softwares

Korb and Nicholson's Appendix B: Bayesian network and causal discovery software packages [pdf]
Kevin Murphy's list of software packages [link]
Java
- JavaBayes [link]
- UnBBayes [link]
Python
- Python Bayesian Network Toolbox [link]
- bayesnet [link]
Matlab/Octave
- Probabilistic modeling toolkit for Matlab/Octave [link] [pmtk data sets]
- Bayes Net Toolbox for Matlab [link]
The instructor's demo implementation of the messaging passing algorithms: SCBayesNet.py, bn-cancer.py, bn-earthequake.py

Syllabus

	Lectures	Assignments
Sep 01	Introduction [slides] Concepts and challenges Various paradigms in expert systems Rule-based systems Bayesian networks	[hw1]
Sep 08	No class
Sep 15	Knowledge representation and methods of inference [slides]	hw2 [discussion]
Sep 22	Probability in AI [slides] Probability and conditional probability Independence Bayesian rules Bayesian views (in comparison with frequentism and propensity interpretation) Utility theories and decision making	hw3 [discussion]
Sep 29	Bayesian Networks & Project Assignments [slides]	[project assignment]
Oct 06	Inference in Bayesian Networks (I) [review] [slides]
Oct 13	Inference in Bayesian Networks (II) [slides] A demo messaging passing implementation in Python: SCBayesNet.py, bn-cancer.py, bn-earthequake.py	Problem 1 of Chapter 3 in Korb& Nicholson's book.
Oct 20	Inference in Bayesian Networks (III) - Junction Tree Algorithms [slides]
Oct 27	Mid-term Exam
Nov 03	Learning in Bayesian Networks [slides]
Nov 10	Decision Networks [slides]
Nov 17	Knowledge Engineering with Bayesian Network [slides]	[hw5] [Matilda.pdf] [software: run on Windows XP or Linux/Wine]
Nov 24	Applications of Bayesian Networks [slides]
Dec 01	Other formalisms of uncertainty reasoning [slides] Default logic Certainty factor Dempster-Shafer theory Fuzzy set
Dec 08	Project presentations [review2]
Dec 15	Final Exam
Dec 22

Grades and Policy

Your grade will be composed of the following:

Homework: 1/3
Mid-term exam and term project: 1/3
Final exam: 1/3

Please hand in your homework before each class meeting. Please acknowledge all sources, including those in the class from whom you obtain the ideas. Late homework will not be accepted.

Yuqing Tang, 2010