Proposal:

--------------------------------------------------------------------------------

Problem:

Many Artificial intelligence programs today work on the principle that all the knowledge in them are preprogramed. Information is not learned but imposed upon them. Even in neural networks the right ansewrs must be known before training can take place and unsupervised neural networks do not always produce the right results that can be used in a meaningful way. Then there is the fact that real world problems such as vocal comunication between man and machine can not be fully implemented by neural networks alone. They must have an expert system behind them doing all the logic.

--------------------------------------------------------------------------------

Purpose:

The purpose of this project is to create a type of neural network capable of preforming all the tasks of traditional A.I. program without the burden of a cumbersome expert system.

--------------------------------------------------------------------------------

Procedure:

To create our neural network we will use the programing language Java. The neural netwok will use hierarchy and sequencial learning technics. The first test data we will use to test our program will be continuous sound running at 200 Hz. Then we will use more complicated sound patterns later on.

--------------------------------------------------------------------------------

Team Members:

Jeremy Wilson

Sponsoring Teacher: Marez

--------------------------------------------------------------------------------

Team Members:

Jeremy Wilson

Sponsoring Teacher: Gregory Marez

Mail the entire Team