University of Houston
Department of Computer Science

In partial fullfillment of the Requirements for the Degree of
Master of Science


Waseem Ahmed
will defend his thesis

An Object Oriented Multi-Dimensional Data Type Learning System

Abstract

Machine Learning is a sub-area of artificial intelligence concerned with the induction of knowledge or refinement of skill. It has attracted a lot of attention since the mid-1950's. Computer programs that learn are called Machine Learning programs. These programs develop concepts, infer new concepts from existing concepts, and revise incorrect concepts. They have been used to create and maintain knowledge bases for many applications.

Presented herewith is a sub-system of a learning program that concentrates on learning multi-dimensional data types, specifically vector and matrix, and operators based on these data types. The system has been designed to learn the multi-dimensional data types (MDDT) based on basic scalar data types. To avoid lengthy and non-understandable sequence of learning steps, several powerful high-level methods have been developed to facilitate the learning of operators for these MDDTs. The learning of the MDDT operators is based on these basic methods and other basic scalar operators. An interface has been provided to demonstrate that these operators have been learned correctly. Through this interface the user of the system is allowed to create multi-dimensional data, and apply the learned operators on the created data.

The system, designed using the object oriented paradigm and implemented in C++, performs consistency checks while learning and applying these MDDT operators. It also performs consistency checks while creating multi-dimensional data.


Date: Wednesday, November 24, 1999
Time: 9:00 AM
Place: 550-PGH


Faculty, students, and the general public are invited
Thesis Advisor: Dr. Kam-Hoi Cheng