Polyhedrons and Perceptrons

From Noisebridge
Jump to navigation Jump to search

Some DreamTeam participants are investigating and possibly experimenting with this unique way of using neural network for solving AI problems.

Polyhedrons and Perceptrons Are Functionally Equivalent, by Daniel Crespin, Facultad de Ciencias, Universidad Central de Venezuela, 05/11/2013

Example code (C++ / MS Windows) and training data developed by author of above paper:

Software events were as follows: I coded Neurogon25.cpp; my son Elias Crespin came in, added more code to create corte.cpp from which .dll and .ocx were obtained providing interface with a GUI.

Source Code:

Media:Poliedros.dat (~54 KB) Read notes stored with file for instructions on making this into a .zip file.

Data Files:

These can be directly fed to neurogon *made from source code).

Media:Sonar.txt (~93 KB)

Media:xor_dim_02.txt (~684 B)

Media:xor_dim_05.txt (~2 KB)

A list of other research reports by Professor Crespin with links and references:

http://www.matematica.ciens.ucv.ve/dcrespin/Pub/CrespinEng.html

http://www.matematica.ciens.ucv.ve/dcrespin/Pub/Crespin.zip

Note: NoiseBridge does not accept .zip files in its wiki,nor does it accept .dat files that are also .txt files.