Previous: atree
Up: NEURAL NETWORK SIMULATION TOOLS
Next: axon
Previous Page: atree
Next Page: axon

AUTONET

General Information
Name
AUTONET
Source
Peak Software Corporation
Price
$ 395
Target usage

Model

Network Architecture
fixed: each neuron in each layer is connected to two neurons in the previous layer, output of it is a square function of inputs. One neuron of the last layer is selected for ourput. The network is generated automatically, the user has no control over the network generation.
Paradigms
Limitations
150 or fewer tupels of input data with a maximum of 24 input variables. network type fixed (see above). number of nodes and number of connections limited by DOS memory (640k)

Implementation

Display
topology of the network, values of units / link parameters, current status, progress in learning, memory usage, ...
User Interface
Menues
Execution
interactive (menue driven) save / restore state (weights) possible. no debugging.
Data Format
ASCII, data types: input is floating point vector, output is floating point scalar. No compression or encoding possible.
Evaluation
graphics, statistics no interface to benchmarks
Performance
use of resources: uses the whole PC, math co--processor recommended.

System Requirements

Hardware
PC math co-processor recomended
Operating System
DOS
Display
color, EGA
Language

Installation and Documentation

Installation
executable
Sources available
Implementation language
Documentation
manual, examples
Support
phone help line
User Groups

franz@neuro.informatik.uni-ulm.de
Wed Jan 19 23:46:24 MET 1994