ANN has been developed on MS-DOS computers primarily for educational uses. Currently, it consists of six simulation programs. ANN1 is a very simple neural net which shows how a network learns by adjusting its connection weights. ANN2 is a single processing element neural net, in which the user trains the network manually by adjusting the connection weights and the threshold value. ANN3 is a manually trained simple two layered network. It demonstrates the power of hidden neurons. ANN4is a Bidirectional Associative Memory network. ANN5 is a Perceptron that learns from examples. ANN6 is a network based on the backpropagation of error. Graphics have been used extensively in all networks. Students can observe the way these networks learn. Hypertext is used to explain concepts, and also serves as an online user's manual.
Malasri, Siripong and Franklin, Stanley P.
"ANN: A Set of Educational Neural Net Simulators,"
Journal of the Arkansas Academy of Science: Vol. 45
, Article 17.
Available at: https://scholarworks.uark.edu/jaas/vol45/iss1/17