Neural Nets - Examples and Simulations 
Sonar returns







This neural networks is used to classify sonar returns. Are the sonar returns coming from a target or simply noise?  For applications such as these, a neural network can often be trained to make more accurate identification and do it more quickly than a human operator

Neural networks tend to excel at pattern recognition far better than more traditional computer programs.  Here's some examples to play with:

Recognize tree species by leaf shape.

Impersonate Spaceman Spiff and train a neural net to land your spacecraft without crashing.  Fortunately, it's only a virtual spacecraft, which is a good thing because it crashes a lot before it gets trained.

Recognize handwriting (well, really it's about drawing letters with a mouse!) with this demonstration applet.  Along the same line, this applet shows how a neural net learns to do OCR (Optical Character Recognition).
 
So how does a neural net learn to recognize a shape?  This demonstration is quite enlightening.

See how a neural network adjusts the weights of its connections between nodes.

Some grab-bag collections of applets--some quite sophisticated--can be found at these websites:  [ 1
| 2 ]

     Home   |   Overview   |   History    |   Tutorials   |   Multimedia and Lectures    
Examples and Simulations   
|   Advanced Topics