CSCI-8050
Knowledge-Based Systems
Prerequisites: (CSCI-6540 and CSCI-6550)
or POD.
Description: (Themes:
Knowledge & Expertise, AI/DB Integration, Decision
Support)
Theory
and practice of knowledge based system construction with particular emphasis on
rule-based expert systems. Topics
include KBS fundamentals, knowledge representation, knowledge base
construction, knowledge integration in databases, inference engines, reasoning
from incomplete or uncertain information, intelligent decision support, and user
tools & interfaces.
Instructor: Walter D. Potter
Office: GSRC-113 (enter through 111)
Phone: 542-0361 (with rollover and voice mail)
Email: potter@uga.edu
Hours: By Appointment, Drop In, or __(hours to be determined)__
Notes: If you stop-by or call, and I’m NOT available
then be sure to leave a note. I’ll get
back to you as soon as possible. E-mail
is best.
Texts (required):
1) Introduction to Expert Systems, Third
Edition by Peter Jackson
2)
Microsoft's Age of Mythology
(tentative)
References
(in Library):
Intelligent Database Systems by Bertino,
Rule-Based
Expert Systems
by Buchanan, B.G. and E.H. Shortliffe, eds.,
PROLOG
Programming in Depth by
Reserve
Books, and Current Literature
Grading*:
|
ES/IDB/IIS |
60% |
Systems,
reports, & presentations (variable due dates) |
|
Assignments |
20% |
Talks,
summaries & other HW (due weekly) |
|
Final
Exam |
20% |
around
Tuesday December 14th, noon |
*No late coursework accepted. Due dates are scheduled in advance and are
firm.
*Class attendance is required and class
participation is graded (under assignments).
Policies: Note that each student is
expected to do his/her own work. Any
evidence of academic dishonesty will not be tolerated and will be subject to
disciplinary action. Be sure you are
familiar with the University’s academic (dis)honesty
policy as well as any departmental policies (see attached). No make-up exams are given.
NOTE: The course syllabus provides a general guide for the course; deviations may be necessary.
CSCI-8050: Knowledge Based Systems
Scope: The
road map we plan to follow this semester includes a focus on three distinct
areas of knowledge based systems: expert systems, intelligent database systems,
and intelligent information systems.
Expert Systems are knowledge
based systems that attempt to rival the performance of a human expert. Typically, a knowledge engineering task is
undertaken to acquire expert domain knowledge from one or more human experts. This knowledge is coded using some useful
representation scheme and possibly some expert system shell IDE. We will investigate the development of such a
system (to the extent allowed within our time constraints).
Intelligent Database Systems
integrate concepts from AI with those from the DB arena to form database
systems with more capabilities than merely serving up facts to user
queries. Active Databases may be
considered a part of the Intelligent Database Systems domain since they use
active triggers (i.e., rules) to initiate some internal database
processing. Other types of rules may be
incorporated into a database to derive values to "virtual" attributes
during query processing. On another
front, rules may be used to massage a user query in order to provide summary
results instead of some large amount of tabular data.
Intelligent Information
Systems bring together several types of systems to help with the decision
making process. A typical IIS has
several components including one or more databases, one or more expert systems,
a structured interface, an intermediate working area (sometimes called a
blackboard), one or more models that can be used for decision making or query
response (i.e., using a forest regeneration simulation model to predict timber
density at some point in the future), and its own processing routines. These components work together in a
transparent fashion to aid user decision making. The infrastructure to support this seamless
interaction among components is the real heart of an IIS.
(Each major topic item is
covered at the approximate rate indicated.
However, due to the dynamic nature of the in-class activities, there may
be substantial variation from this schedule.)
Week 1 Expert Systems - Introduction
Definition
Characteristics
Typical
Applications
Example
Systems
Week 4 Components of Expert Systems (Architecture)
Knowledge
Base
Knowledge
Representation
Meta-Knowledge
Inference
Engine
Search
Techniques
Reasoning
With Uncertainty
User
Interface
User
Dialog
Explanation
Tutoring
Week 9 Tools and Environments for Expert System Development
Week 10 Building an Expert System
Problem
Selection
Development
Methodology
Knowledge
Acquisition
Pitfalls
Week 12 Evaluation of Expert Systems
Test
Cases
Refinement
Performance
Week
14 Intelligent Database Systems
Data Models
Active
Database Systems
Derivable
Attribute Values
Week
17 Intelligent Information
Systems
Blackboard
Architecture
Wrapper
Architecture
Dependent
Agent Architecture