Contents
This course will cover major areas in the field of Artificial Intelligence,
including machine learning, neural networks, genetic algorithms, planning,
computer vision, and reasoning about uncertainty. Several programming
assignments will provide practice with AI techniques. Students will also take
two exams and complete a semester project of their own design.
- Time: Monday, Wednesay, and Friday, 10:00 - 10:50 AM
- Place: EDUC 105B
Knowledge of fundamentals of artificial intelligence and logic. Some
familiarity with Common Lisp is also recommended. If you have not taken
CS440 but have taken an introductory artificial intelligence course elsewhere,
please see the instructor before registering for this class.
- Software:
Access to a Common Lisp interpreter is required. All registered students
have access to our Allegro Common Lisp environment. You may also obtain
Common Lisp for your own computers. There are a number of commercial and
public-domain Common Lisps available. See Resources
below for more information. The assignment involving UCPOP will require
Common Lisp. For other assigments, you are free to choose the language.
- Homework: Three programming assignments will be given.
These will require implementation of a technique, its application to a
problem, and a well-written report on what was done for the assignment. The
report must also include extensive interpretation of the results and
speculation on how variations in the problem and technique would affect the
results.
- Exams: A midterm and final exam will be given. The
final will not be a comprehensive exam.
- Semester Project: The semester project provides the
students with the opportunity to specialize in an area of AI that is of
interest to them. A project might consist of applying a publicly-available AI
system to a hard problem, a review of several AI research articles,
implementations of AI techniques other than those covered in the assignments,
or some combination of these.
Projects must be discussed with the instructor before March 2nd. On March
2nd, a written proposal must be handed in at class time and the proposal must
be presented in class. On April 6th, all students will present a short
progress report in class. Written project reports are due in class on May
6th. On May 6th and 8th, the projects will be presented in class. The
project grade will be determined by the proposal, the progress report, and the
final report, and in-class presentations. So, all phases of the project will
be graded.
- Late Assignments and Working Together: You are strongly
encouraged to discuss assignments with each other, but you must write your own
code and report. It is possible for several students to work together on a
semester project, but this must be approved by the instructor before the
proposal date.
Late assignments will be accepted, but the grade will be reduced by 5\% for
every day past the due date, for a maximum of 25\%.
Be sure you read the Computer Science Department's
Student Information Sheet. It explains the department's policies
regarding late assignments and other important things.
Your grade will be based on homework assignments, two exams, and a semester
project, weighted as follows:
- 30% for homework assignments,
- 15% for midterm exam,
- 15% for final exam,
- 40% for semester project.
- Jan. 21: Overview.
- Jan. 23: Neural networks. The What and the Why. Paper: "The Appeal of Parallel Distributed Processing"
- Jan. 26: Training neural nets. Paper: "Learning Internal Representations by Error Propagation". Text: 19.1 - 19.5.
- Jan. 28: continued.
- Jan. 30: continued.
- Feb. 2: continued.
- Feb. 4: Bagging and Boosting. Paper:
"Machine Learning Research: Four Current directions", by T. Dietterich. A
postscript
preprint is available.
- Feb. 6: continued
- Feb. 9: Making Complex Decisions. Text: 17.1 - 17.3.
- Feb. 11: continued
- Feb. 13: Reinforcement learning. Text: 20.1 - 20.7. Paper:
Dietterich's paper. See above link.
- Feb. 16: continued
- Feb. 18: continued
- Feb. 20: continued
- Feb. 23: Uncertainty. Text: 14
- Feb. 25: continued
- Feb. 27: continued
- Mar. 2: Project proposals presented in class.
- Mar. 4: Review for Midterm Exam. See this topics list.
- Mar. 6: Midterm Exam
Mar. 9-13: Spring Break
- Mar. 16: Genetic algorithms. Text: 20.8. Papers: "A Genetic
Algorithm Tutorial", by Darrell Whitley. "Artificial Intelligence: Theory and Practice", by Dean, Allen,
and Aloimonos.
We will also discuss
other search methods described in our Text: 4.4
- Mar. 18: continued
- Mar. 20: continued
- Mar. 23: Probabilistic reasoning systems. Text: 15
- Mar. 25: continued
- Mar. 27: Making decisions. Text: 16. Making complex decisions. Text: 17.
- Mar. 30: continued
- Apr. 1: Machine Learning. Text: 18
- Apr. 3: continued. Paper: "Machine-Learning Research"
- Apr. 6: Project progress presentations in class.
- Apr. 8: Machine learning continued
- Apr. 10: continued
- Apr. 13: continued
- Apr. 15: Planning. Text: 11-13
- Apr. 17: continued
- Apr. 20: continued
- Apr. 22: continued
- Apr. 24: continued
- Apr. 27: Computer Vision. Text: 24.1 - 24.6
- Apr. 29: continued
- May 1: continued
- May 4: continued
- May 6: Semester project presentations in class.
- 10:00 Laura
- 10:15 Tom
- 10:30 Xichen
- 10:45 Sridhar
- May 8: Semester project presentations in class.
- 10:00 Kae
- 10:15 Josef/Alex
- 10:45 Michael
- May 11: Semester project presentations in CS conference room.
- 2:00 Charlie
- 2:15 Brian
- 2:30 Paul
- 2:45 Doug
- 3:00 David
- 3:15 Youbin
- May 12: Final Exam, 1:30 - 3:30 PM, same class room. Topics for final exam.
List of Assignments:
- Assignment 1, Due February 16th (March
2 for NTU students). Here are some comments for everyone about what you turned
in for this assignment. Please read these comments in preparation for
the second assignment.
- Assignment 2, on genetic algorithms. Due April 1st (April 15 for NTU and SURGE students).
- Assignment 3, on planning with UCPOP. Due April 24th (May 1 for NTU and SURGE students).
- For help with your writing, see the CSU Writing
Center web site for tutorials and descriptions on improving your writing.
You may even submit drafts of your papers and receive editorial feedback from
the Writing Center's consultants! This is an extremely valuable
resource!
- Belief Networks
- Reinforcement Learning
- Applet that learns to play BlackJack using reinforcement learning,
- Matlab
- Artificial
Intelligence: A Modern Approach, our text book author's site that
includes very useful links to on-line AI resources,
- AI Resources,
maintained by the Intitute for Information Technology in Canada,
- Common Lisp
- Common Lisp - The
Language , 2nd Ed. G. Steele, Digital Press, 1990
- Franz Inc., where you download a
version Common Lisp for Windows that does not include a compiler,
- CMUCL, a Common Lisp maintained
by CMU for Unix
- UCPOP
- Newsgroups for artificial intelligence include
comp.ai and
others focused on particular subfields, such as
comp.ai.neural-nets, comp.ai.genetic
comp.ai.vision, comp.ai.nat-lang,
comp.ai.shells, comp.ai.fuzzy,
comp.ai.alife. And don't forget
comp.ai.jair.announce that announces new papers added to the
electronic Journal of AI Research.
Due dates for SURGE and NTU students are two weeks later than the regular due
dates. The project grade for SURGE and NTU students will be based solely on
their written work. However, all students are welcome to attend and
contribute to the in-class presentations.
Copyright © 1998 Chuck
Anderson
anderson@cs.colostate.edu
CS540: Artificial Intelligence (Spring, 98)