Syllabus
UNIVERSITY OF DAYTON
CPS 481: Advanced Artificial Intelligence
Winter Semester 2008 3 credits
Meets: M, W 3:00pm – 4:15pm
203 Miriam Hall
|
Professor: Dr. Jennifer Seitzer Office: 144 Anderson Hall Email: seitzer@udayton.edu Course Web Page: http://homepages.udayton.edu/~seitzer/cps437 Phone: (937) 229-2197 *Office Hours · M, W 4:15 – 5 pm · Friday 3:00 –4:00 pm · BY APPOINTMENT --------------- *These are tentative and will possibly change due to departmental meetings and other college events as they arise. |
Mailing Address Dr. Jennifer Seitzer, Associate Professor Computer Science Department University of Dayton 300 College Park Dayton, OH 45469-2160 |
Prerequisites:
CPS 480: Artificial Intelligence.
Motivation:
Artificial Intelligence is the sub-discipline of computer science that attempts to endow intelligence to computer systems. Although AI has produced some very sophisticated computer systems, it is generally accepted that the goal of systems achieving human-level intelligence is still elusive.
There are two general approaches to the pursuit of artificial intelligence. The symbolic is often referred to as "good old fashioned AI" (philosopher John Haugeland) and is abbreviated as: GOFAI. This is based on the belief that thinking can be accomplished by a physical symbol system -- that the manipulation of a physical symbol system provides necessary and sufficient means for intelligent behavior [Newell and Simon 1976]. The subsymbolic approach to AI employs techniques that mimic organisms in nature including the human neuron as well collective and individual insect behavior. In fact, part of this camp of researchers believe that thinking can only occur in machines made of proteins [Searle 1980, 1992].
In this course, we will address some traditional areas of AI including default reasoning, fuzzy logic, planning, and machine learning. We will also broach a few topics spawned from the subsymbolic approach including neural networks and perception.
Objectives:
· To understand some non-classical logics including nonmonotonic reasoning, fuzzy logic, and probabilistic reasoning with uncertainty
· To program an expert system using a preexistent expert system shell and your own unique knowledge rule base.
· To grasp the dominant areas of machine learning including version space, inductive logic programming, genetic algorithms, clustering, and neural networks
· To study some of the fundamentals of natural language processing
· To understand the earliest algorithm of Swarm Programming: shortest path mimicking ants
·
To program the techniques and
algorithms studied in class into working systems
Subject Matter (Tentative list and schedule of coverage):
|
Wk |
Monday Lecture |
Wednesday Lecture |
|
1--
|
1/07/2008: Course Introduction; Review of Propositional and First Order Logic, and PROLOG |
1/09/2008: Chapter 10.7 More Prolog |
|
2-- |
1/14/2008: Chapter 10.7 Introduction to Nonmonotonic Reasoning: Default Reasoning |
1/16/2008: Default Reasoning; Truth Maintenance Systems; Introduction to Stable Models and Negation-as-Failure |
|
3-- |
1/21/2008 · Demonstration of How to Present a Paper: Paper on Nonmonotonic Reasoning and Stable Models · Ideas on Implementation of a JTMS |
1/23/2008 · Introduction to Fuzzy Sets and Fuzzy Logic · Fuzzy Logic: The Cement Kiln Problem |
|
4-- |
1/28/2008 · Presentation of a paper on Default Reasoning/Nonmonotonic Reasoning · Fuzzy Logic: The Cement Kiln Problem NOTE: MON., 1/28/2008 Last day to Withdraw without record |
1/30/2008 · Presentation of a paper on Default Reasoning/Nonmonotonic Reasoning · Fuzzy Logic: Building and Using Fuzzy Controllers |
|
5-- |
2/04/2008 · Presentation of papers on Fuzzy Logic · Review for Test 1
|
1/30/2008 Test 1 on material covered so far
|
|
6-- |
2/11/2008
|
2/13/2008
|
|
7-- |
2/18/2008 · Presentation of Paper on Expert Systems · Computational Induction · Introduction to Knowledge Discovery
|
2/20/2008
|
|
8-- |
2/25/2008 Machine Learning: Inductive Logic Programming |
2/27/2008 Machine Learning: More Inductive Logic Programming |
|
9-- |
3/03/2008 · Presentation of Paper on Inductive Logic Programming · Introduction to Genetic Algorithms |
3/05/2008 · Machine Learning: Genetic Algorithms
|
|
10- |
3/10/2008 · Machine Learning: Genetic Algorithms · Presentation of Paper Genetic Algorithms
|
3/12/2008 · Presentation of Paper Genetic Algorithms · Introduction to Neural Networks
|
|
11- |
Spring Break: no class |
No class |
|
12- |
3/24/2008 Spring Break: no class |
3/26/2008 Neural Networks (continued) |
|
12- |
3/31/2008 · Test 2
|
4/2/2008 · Presentation of Papers on Neural Networks · Presentation of Paper on any topic of Machine Learning · Communication: The Underpinnings of Natural Language Understanding (NLU) ---Chap 22
|
|
13 - |
4/07/2008 Natural Language Understanding (continued)
|
4/09/2008 · Presentation of Paper on NLU · Intro to Information Retrieval (IR) · Relationship of NLU and IR
|
|
14- |
4/14/2008 Introduction to Web Mining Use of IR techniques in Mining the Web |
4/16/2008 · Presentation of Paper on Web Mining · Introduction to Swarm Intelligence
|
|
15 - |
4/21/2008 · Presentation of Paper on Web Mining · Introduction to Swarm Intelligence
|
4/23/2008 · Introduction to Swarm Intelligence · Paper on Swarm Intelligence
|
|
16 - |
4/28/2008 Finals Week Time to be announced |
|
Required Text: Artificial Intelligence A Modern Approach
By, Stuart Russell and Peter Norvig
ISBN # 0-13-103805-2
Grading (Approximate distribution of credit):
Test 1 30 points
Test 2 30 points
Final Exam 45 points
Paper Presentation 30 points
Graduate Student Project 30 points
In-Class Grade 10 points
(includes
attendance, class participation, and in-class exercises)
* note: these cannot be made up
Grading Scale – The new university scale will be used for indicated letter grades.
Undergraduate
Grade Score
A Excellent 93 +
A- 90-92
B+ 87-89
B Good 83-86
B- 80-82
C+ 77-79
C Fair 73-76
C- 70-72
D Poor/Passing 60-69
F Failing <60
Policy on Makeups, Missed, and Late Work:
1. Late
Work: Work will usually be accepted late and recorded as such. Work is due
at the beginning of class. A 10% penalty is applied for every class day
the assignment is late. No work will be accepted after the assignment has
been graded and returned, or after solutions have been given out in class.
2. Make-ups:
Tests are expected to be taken on the test date. Any make-ups must be
established with me ahead of time. There are no make-ups for in-class chapter
quizzes, exercises, or participation. To allow for one unavoidable quiz
absence, I will drop the lowest quiz grade. In general, however, to get these
points, you must come to class.
3. Attendance: Students are expected to come to class. If a class must be missed, however, students are responsible for all material, assignments, and announcements made during class. For this reason, you are encouraged to find a colleague with whom you can communicate to share such important information. Pop quizzes and in-class work may not be made up. To get the points, you must come to class.
Email Communication and Class Computer Accounts:
Email: I prefer to conduct communication through email. My email address (as indicated above) is seitzer@.udayton.edu. Please feel free to write me anytime. I try to check my email many times through the day. If you do not have an email account, I ask that you get one. Student email accounts can be acquired from the Systems Administrator. For information, you may call (937)229-3858.
Course Web Pages:
· The course has its own web page that can be found at URL http://homepages.udayton.edu/~seitzer/cps481. You are responsible for consulting this page regularly. Most handouts and other communications will be posted on this page. Additionally, the textbook has a web page. Many slides used in class can be procured from the textbook site that can be found at URL http://www.cs.berkeley.edu/~russell/aima.html .
Class Email List:
· Along with web page postings, I regularly send my classes email via the respective Class Email List. Please make sure you have the correct address logged with the university to received all class emails. These lists are maintained by the university.
Lab Conduct Policy:
A list of rules is posted in the Computer Science labs. Please read these rules and remember that these labs are shared spaces and that we need to be considerate of all lab users. Only printing of Computer Science course materials is permitted in the labs.
Graduate Student Project
Graduate students are required to produce a final project for the course. This semester, each graduate student should choose a topic of Artificial Intelligence presented in class and create an implementation that captures some of the main concepts of that topic. In particular, the following should be produced and submitted:
1. a one paragraph description of the project to be approved by instructor about eight weeks into the course. This should be approved by me before proceeding to the next step.
2. a requirements/design document describing the behavior and design of the system to be submitted three weeks before due date
3. a well-documented program
4. the Powerpoint slides which should be presented to the class in a 15 minute presentation (including a system demo)
5. a two-page (double-spaced) paper describing the system.