Nice simulator of neural networks which has been developed at the university of Stuttgart. All parameters are presented in individual windows. To really use this program a fundamental background of neural networks is required.
SNNS (Stuttgart Neural Network Simulator) is a software simulator for neural networks on Unix workstations developed on the university of Stuttgart. The aim of the project is the development of an flexible simulation interface for research and usage of neural networks.
SNNS consists mainly of two parts:
o simulator kernel written in C
o GUI under X11
The simulator kernel works with the internal network-data structures of the neural network and performs all operations by learning or doing. It can also be used separately as a C program embedded in own applications. It supports multiple network topologies and almost like RCS supports the sites concept. SNNS can be extended by the individual user with user-specific activation functions, output functions, site-functions and learning procedures which can easily be connected as simple C programs with the simulation kernel.
At the moment of this writing the following algorithms are supported:
o back propagation
o vanilla (online) BP
o BP with Momentum Term and Flat Spot Elimination
o Batch BP
o Time Delay BP
o Counter propagation
o Quick prop
o Back percolation 1
o Generalized radial-base-functions (RBF)
o RProp
o ART 1
o ART 2
o ARTMAP
o Cascade Correlation
o Return Cascade Correlation
o Dynamic LVQ
o Back propagation through time (for recursive networks)
o Batch Back propagation through time (for recursive networks)
o Quick propagation through time " " "
o Kohonen Feature Maps
The GUI XGUI based on the kernel presents you 1/2/3-dimensional views of the neural network and controls the kernel during the runtime of the simulation. Additionally the 2D user-interface provides an integrated network-editor, which can be used to design, manipulate and present neural networks.
A complete user manual can be found in the package snnsdoc.
The current version of SNNS is to be obtained from ftp.informatik.uni-stuttgart.de /pub/SNNS