UNIVERSITY OF GUELPH

Department of Computing & Information Science

COURSE OUTLINE

Course Code:
CIS*6020 / Course Title:Artificial Intelligence / Date of Offering:
Fall 2008
Instructor:Dr. Yang Xiang
Office hours: Mon, Wed, 3:30-4:20 Pm / Office:
Extension: / Reynolds 318
52824
Email: /
Calendar Description:
An examination of Artificial Intelligence principles and techniques such as: logic and rule based systems; forward and backward chaining; frames, scripts, semantic nets and the object-oriented approach; the evaluation of intelligent systems and knowledge acquisition. A sizeable project is required and applications in other areas are encouraged.
Prerequisite:This course is intended for entry graduate students who may or may not have prior undergrad AI background. For students without prior AI background, supplementary readings are required and will be assigned early in the course.
/ Topics:
  • Agents and environment, properties of environment, search, knowledge-based agents, environment – representation – inference, logic inference, and inductive learning.
  • Constraint satisfaction: constraint satisfaction problems, backtracking, iterative improvement, and constraint propagation.
  • Probabilistic inference: limitation of logic, notation and axioms of Bayesian probability, inference with JPD, Bayesian networks (BNs), inference in BNs by variable elimination, and building models as BNs.
  • Machine learning: ensemble learning, computational learning theory, learning Bayesian networks, support vector machines, and reinforcement learning.

Tentative Schedule:
Lecture Tue, 2:00-4:50 Pm, Reyn312
Main Text:S.Russell and P.Norvig,Artificial Intelligence: A Modern Approach, (2nd Ed.),
Prentice Hall, 2003.
Method of Evaluation:
Course Work / Date / Weight
Assignment 1:
Assignment 2:
Assignment 3:
Project: / Weeks 4 to 6
Weeks 8 to 10
Weeks 11 to 12
Weeks 8 to 12 / 10%
10%
10%
20%
Final Exam / TBA / 50%
Grading Policies / The student must attempt all course work and in at least 70% of these elements must achieve a passing grade.
Website:

ACADEMIC INTEGRITY

The Univ. of Guelph is committed to the highest standards of academic integrity and honesty. Students are expected to be familiar with these standards, and must abide by the applicable policies (see Section VIII of the Undergraduate Calendar on "Academic Misconduct").

For educational purposes, instructors impose conditions on assignments that may limit students' permission to collaborate with others or to utilize external sources (including, but not limited to, software, data, images, text, etc.). Any permitted utilization must be done with proper references. Instructors may use automated tools to detect possible cases of plagiarism. Work that shows significant unnatural similarity, or that appears to be copied from unacknowledged sources, will be investigated as potential academic misconduct. "Aiding and abetting" is also a punishable offence, and students must be careful not to help others commit offences by giving out their files or allowing others to access their computer accounts. Consider yourself warned.

Learning from each other among students is encouraged through discussion on concepts and algorithmic ideas. However, coding of programming assignments and phrasing of answers to short assignment questions must be completed independently by each student, unless the assignments involved are specified to be completed by team work.

ACCEPTABLE USE POLICY FOR COMPUTING FACILITIES

Please read the complete Univ. ofGuelph policy at

CHANGES IN DATES ON COURSE OUTLINES

See: Undergraduate Calendar: VIII. Undergraduate Degree Regulations and Procedures

Grading Procedures (Resolution 5)

E-MAIL POLICY

Students should include their name and course number in every email, e.g. Joe Smith: CIS*xxxx, since instructors are often involved in teaching more than one course per term. To comply with university privacy policy, all emails should be sent from your uoguelph account (not from hotmail.com, gmail.com, or any other non-UoG host). All students are responsible for reading their uoguelph email and therefore should maintain their accounts, i.e. disk quotes should be monitored so that email is not rejected due to lack of space.

ASSIGNMENT SUBMISSION

Assignments are expected to be handed in by specified due dates. Late assignments will be subject to 15% of the total mark per day up to 3 calendar days (not weekdays). For instance, the latest date to hand in an assignment due on a Friday is the following Monday, with the highest possible mark of 55%. The submission procedure is detailed in the course website.

The first page of each assignment is required to clearly indicate your name, student id, course number and assignment number. Keep a copy of all your submitted assignments in case they get misplaced. Keep your returned assignments/midterm. In the rare case your grades get recorded incorrectly, use the returned copy to get the problem resolved.