Welcome to my pre pre pre beta version of a Java Back Propagation Network simulator. As you can see by the extrodinary amount of pre's preceding the word beta, this is in a very experimental stage.
This network is designed to learn the logical OR function. For those of you who don't know, OR is a logical function that has two binary inputs and one binary output. If the either one of the input values is one, then the output value is one. If both of the input values are zero, then the output value is zero.
For the current version of the program, you can only train the network and watch the aggregate (total) error decline on the graph. To begin, click on Add Patterns. Then click on Train. If you wish to interrupt the training process press stop. If you want to start over, press Reset Network. The two inputs at the bottom are for the learning constant and the momentum constant. Fill in those values before you start training. You can change them in the middle of the training process by stopping the training, changing the values, and then starting the training process again (do not click reset network).
That's all the instructions you need for now so give it a try.
If you would like to see the source code for this program click here.