Making computers intelligent

October 3, 2011

By Brigitte Graf ’13

Dr. Christos Nikolopoulos, professor of computer science and information systems, has received an external research grant from Caterpillar in the amount of $50,000. The duration of the grant will be for one year, through May 2012.

This grant will support the research of Dr. Nikolopoulos, three graduate student assistants and a few faculty members who will be developing "Machine Learning Algorithms for Demand Forecasting.”

Demand forecasting involves the prediction of what the demand for a particular product will be so a company is able to adequately supply that product. The goal of this project is to invent machine learning techniques which would enable a computer to work out these predictions without a programmer.

“A computer, in a sense, is a very stupid machine,” Nikolopoulos said. “We give instructions, and all the computer does is execute them; in essence, we are making the hardware intelligent.”

According to Nikolopoulos, there are several ways to approach demand forecasting. The one he intends to take is of a more statistical nature.

“Based on the data, we develop heuristics that the computer can learn itself so it can detect patterns and discover knowledge embedded in the whole data set,” Nikolopoulos said.

The benefits will be numerous for the students working with Dr. Nikolopulos on this project. Should the project be executed successfully, they could be involved in publications and patents that result from the research.

The most significant of these advantages, however, is the opportunity for exposure. Due to the nature of the project, the students involved can hope to gain experience in a professional setting.

“The students are working with a real-world problem,” Nikolopoulos said. “They are working with professional engineers, so the benefits also include networking.”

Research on this project began a month ago. The grant itself will also generate at least $15,000 in "indirect costs" for Bradley.