This project was started May, 2004 as a senior research project by
Jack Sigan in the Department of Electrical and
Computer Engineering at
Although mainly a research endeavor, the end goal of this project is to produce a system based on artificial neural networks (ANNs) which will play chess (as the dark side only) effectively against a human opponent. The dark side is chosen simply because it prevents the system from having to make the first move. To meet the objective, research will be centered on artificial neural network (ANN) topology specifically for the purpose of creating a topology appropriate for complex decision making in a massively dimensional problem. Functionality of the system may be broken into three parts: the learning mode will be an automated process where the ANNs learn how to play from an extensive external database of games; while the playing and advisory modes of operation accept user interaction, taking move inputs from the player and providing the move(s) chosen by the system. Playing mode returns one move while advisory returns multiple.
This site contains all of the files produced for this project, as well as the laboratory notebooks. Reproduction or redistribution of any material is prohibited. Please see the contact page if you should have any questions or insights related to this project. All files are located on the “downloads” page.