Course Instructor

Name:

Yurai Núñez-Rodríguez

Email:

yurai [at] cs.queensu.ca

Website:

www.cs.queensu.ca/home/yurai

Office hours:

Mondays 4:00pm to 6:00pm, Goodwin 533. For other hours feel free to contact me by email. I am also available for answering questions right after every session or anytime you see me around.

Note: if you come to my office hours after 5pm, you may find the hallway door closed. Please, call my office phone ext: 74671.

Teaching Assistant

Name:

Lili Wang

Email:

lili [at] cs.queensu.ca

Office hours:

Mondays 11:00am to 12:00pm, Goodwin 230.

Course Description

Following a "modern approach" to Artificial Intelligence, we will study the design of agents that act intelligently, that is, agents that act rationally towards the completion of a given task. In order to do so, Artificial Intelligence integrates aspects from different disciplines, such as Logic, Algorithms, Probability Theory, Natural Language Processing, Computer Vision, etc.

One of the main challenges when solving problems for which rationale is required is the typical intractability of the problems: considering all the reasonable possibilities towards achieving the desired goal may involve a number of steps that is exponential in the size of the world under consideration. Many of the problems we will study in this course have no known deterministic polynomial time solution. Also, one may want to design agents that can adapt and keep up with the ever-changing world. The implementation of such flexibility as part of an agent is, in many cases, in contraposition with efficiency. Another difficulty one may encounter when designing an intelligent agent is that the information available to the agent is imprecise, ambiguous, or contains errors.

In order to overcome these difficulties, as best as possible, AI has gathered a set of techniques that allow the use of rationality to solve several types of problems that are commonly found in real-world situations.

For a list of those topics that will be covered see our tentative weekly schedule.

Prerequisites

CISC 352* Artificial Intelligence

Goals

By the end of the course the student should:

• Have state-of-the-art knowledge of what computers can achieve, and how this compares to the set of problems that humans can solve.

• Be able to identify, classify, and model a problem such that a known AI technique or tool can be applied to it in order to "efficiently" find a solution.

• Be able to develop AI tools to solve a certain type of problems.

• Be able to conduct (basic) research, individually or in (small) groups and effectively present his/her results.

• Be one step closer to being an accomplished scientist.

Our approach

• There will be a ~3-hour session every week.

• A set of warming-up questions will be asked at the beginning of each session with the purpose of refreshing on previous topics and introducing the new ones.

• Sessions will be as interactive as possible. This means that student's participation will be key part of the sessions, as well as part of the evaluations (see participation marks and student presentations under assessment and evaluations). Discussions are encouraged. Group work, and brainstorming will also be .

• Most of the new material will be introduced by means of a small sized practical examples/problems.

• During the sessions, we will use PDF slides. More theoretical results will be derived using the blackboard.

Bibliography

Textbook

• S. Russell an P. Norvig, Artificial Intelligence: A Modern Approach (2nd edition), Prentice Hall, 2003. (AIMA2E)

You can buy AIMA2E from the campus bookstore. Notice that, because AIMA2E has been the textbook of choice for CICS-453 for at last the last 3 years, used copies may be available (at a lower price) .

For other material related to the textbook, such as code for some of the algorithms presented on the book and some of the lecture slides, visit the AIMA website.

Other books (that may be used for the course)

• R. Knight, Artificial Intelligence (2nd edition), Mc Graw Hill, 1991.

• P. H. Winston, Artificial Intelligence (3rd edition), Addison-Wellsley, 1992.

Other Resources

• WebCT: for news, class materials, assignments, marks, this syllabus, etc.

• Artificial Intelligence on the WWW:

  • AI on the web also from the AIMA website.

  • www.aaai.org for an update on recent events related to Artificial Intelligence. See for example interesting AI related videos.

  • MIT videos on AI topics.

  • Others: Several contemporary editions of books on Artificial Intelligence are complemented with code, exercises, and lecture notes available online, just find the book/author website and off you go!

Assessment and Evaluations

• Homework exercises will be oriented on a weekly basis. Homework will be part of the student preparation and assessment. The solution of a subset of the oriented exercises will be discussed in class. This will likely happen on the sessions before the exams. Notice that, although no marks are granted directly for homework exercises, these will allow students to participate by presenting solutions of the posted exercises and, therefore, will provide the benefit of participation marks (see below).

• There will be one in-class midterm quiz.

• There will be two assignments for students to complete individually. The solutions to the assignments will be submitted through WebCT -instructions will be provided for this. Late submissions will be penalized at a rate of 20% (of the total assignment value) per day.

• A research project will also be part of the evaluation. The research project can be taken up individually or in pairs. More details on the research projects will be specified later in the term along the list of possible topics.

• In class participation will be rewarded as well. Each participation will account for 0.5% of the total course value, until the 5% limit is reached. Students are allowed a maximum of 3 interventions per session. It is the responsibility of the student who participates in a session to approach me at the end of that session to get his/her corresponding marks recorded.

• There will be one final exam. The final exams will contain a set of optional questions in order to accommodate the students' area of expertise -possibly related to their research project.

Thus, the marking scheme is as follows:

• 15% midterm

• 20% Assignments (10% each)

• 20% Research project (10% presentation, 10% report)

• 5% Participation in classes

• 40% Final exam

Academic integrity

Check out Queen's statement on five fundamental values that characterize academic integrity: honesty, trust, fairness, respect and responsibility (as published online by the Faculty of Arts & Sciences)