NEAT is Neuro-Evolution of Augmenting Topologies, a neuro-evolution technique devised
by Kenneth Stanley, currently of the
University of Central Florida School of Electrical Engineering and Computer Science.
SharpNEAT is a complete implementation of NEAT written in C# and targeting the .Net runtime
environment (running on both MS Windows and Mono/Linux).
Research using SharpNEAT
Note. The papers linked below predominantly use HyperSharpNEAT, a version of SharpNEAT extended by David B. D'Ambrosio at the University of Central Florida to implement 'HyperNEAT' - an extension to the NEAT method that uses a new form of indirect/generative encoding.
2009
-
A Hypercube-Based Indirect Encoding for Evolving Large-Scale Neural Networks, Kenneth O. Stanley, David B. D'Ambrosio, and Jason Gauci
To appear in: Artificial Life journal. Cambridge, MA: MIT Press, 2009 (Manuscript 39 pages)
-
Generative Encoding for Multiagent Learning, David B. D'Ambrosio and Kenneth O. Stanley
In: Proceedings of the Genetic and Evolutionary Computation Conference (GECCO 2008). New York, NY: ACM, 2008 (8 pages)
Winner of the Best Paper Award in Generative and Developmental Systems at GECCO-2008
-
A Novel Generative Encoding for Exploiting Neural Network Sensor and Output Geometry, David B. D'Ambrosio and Kenneth O. Stanley
Nominated for Best Paper Award in Generative and Developmental Systems, GECCO 2007David D'Ambrosio and Kenneth Stanley at the University of Central Florida used their extended version of SharpNEAT - HyperSharpNEAT - in this paper presented at GECCO 2007. This was one of two papers introducing HyperNEAT, a powerful new indirect encoding.
Their HyperSharpNEAT releases are available from the UCF EPLEX software download page.
-
Wesley Tansey extended SharpNEAT to allow parallel processing across multiple CPUs while at Virginia Tech. This work is hosted on Codeplex as ParaSharpNEAT.
SharpNEAT on the web
- Genetic Art with SharpNEAT. Holger Ferstl created this SharpNEAT based application that evolves images by allowing a user to repeatedly select their favourite from a page of candidate images. The application was built on the .Net 2.0 framework and is capable of generating some pretty amazing output. Many thanks to Holger. Now for the eye candy - these were generated by me, there are more examples through the link...
|
|
Experiment/Investigation write ups
- A Neural Network Controller for the Physical Travelling Salesperson Problem. [July 2005]
- A Fast Equivalent for System.Random. Article posted to CodeProject.com [December 2004]
- An Integer Based Neural Network [September 2004]
- Phased Searching with NEAT: Alternating Between Complexification And Simplification. [August 2004]
- Neural Network Code Optimizations [June 2004]