Adaptive and Neural Networks & Tools

CNLS-Net: Dr. Mead codeveloped the Connectionist Normalized Local Spline Neural Network (NN) while on sabbatical in 1991 at the Center for Nonlinear Studies (CNLS) .

Dr. Roger D. Jones initiated and led the effort that created CNLS-Net. Many others contributed to its realization. Dr. Mead joined this outstanding team and helped refine the net into an outstandingly useful NN.

CHC-Net: In 1992-3, Drs. Mead and Jones coinvented the Connectionist Hyperprism Classification Neural Network under the auspices of the DOE and the DOT.

S-Net: Select Net was invented by Dr. Mead to serve two special functions in the Adaptive Arithmetic Tutor.

In addition to these "invented-here" NN's, ANSR has access to and makes appropriate use of many other networks that are widely known in the field, including

• Feedforward-Backpropagation Net (often referred to as the "Back-Propagation Net" or "BP Net")
  
• Adaptive Resonance Theory (ART) Net (Carpenter and Grossberg, 1987)
  
• Learning Vector-Quantization (LVQ) Net (Kohonen, 1987)
  

ANSR sometimes uses commercially-available NN packages, such as

• Brainmaker by California Scientific Software
• NN Toolbox for MatLab by The Math Works, Inc.

Ready to look at some NN application examples?

Return to ANSR Home Page or Site Contents -- or -- Contact ANSR.