Neural-network software

Chapter 2 describes a number of artificial neural-network algorithms used in simulations of psychological development, with some emphasis on cascade-correlation and related techniques. The following links provide access to some of these programs.

*      We wrote a Java applet that runs cascade-correlation networks. No programming experience is necessary – just paste in your training patterns and optional test patterns and click. Direct comparison with back-propagation networks is available as an option. Various canned problems are also included to provide experience with network performance.

*      Code for a variety of neural network algorithms, many of which are discussed in chapter 2

*      Cascade-correlation code in the Lisp and C languages

*      Recurrent cascade-correlation code in the Lisp and C languages

*      Quickprop code in the Lisp, Scheme, and C languages

*      The Stuttgart Neural Network Simulator (SNNS) has a C-language version of cascade-correlation, recurrent cascade-correlation, quickprop, and other algorithms

*      Cascade-correlation and quickprop code in the C language, and other algorithms

*      LNSC simulator package, a full-featured simulation package for cascade-correlation in Lisp