Slashdot Mirror
Open Source Neural Networks?
Sleexy asks: "Does anyone know of
any open source neural network / Artificial
Intelligence projects for any platform,
or anyone doing open source work in this
field?"
← Back to Stories (view on slashdot.org)
← Back to Stories (view on slashdot.org)
I have some code I whipped up a few years ago when I was playing with this stuff. I'll have to dig around for it, but if you email me, I can send it your way.
æeee!
I'm coming up with a relatively AI-y database system that is intended to store and process information like a mind; it's called the Node-Oriented Database Engine (NODE). It'll be open source, and probably written in Scheme or Common Lisp. If you're interested, email me.
To the editors: your English is as bad as your Perl. Please go back to grade school.
I'm coming up with a relatively AI-y database system that is intended to store and process information like a mind; it's called the Node-Oriented Database Engine (NODE). It'll be open source, and probably written in Scheme or Common Lisp. If you're interested, email me.
To the editors: your English is as bad as your Perl. Please go back to grade school.
It was written in C++ 2 years ago has an adaptive network engine and a fuzzy system to adaptive network translator. Documentation is only available in hungarian, but I can give you pointers to the papers which served as a basis for the program.
Drop me a line if you're interested and I'll send the source or provide access to it. On the other hand I have seen many similar things at CMU archive, so you may have luck there.
I got all of these off of Google. Try entering 'neural network' as a search term and seeing what hits you get.
I may have some old notes, papers and source code on them in that area of maximized entropy I call my apartment. If I can find anything I'll post them and GPL everything.
I have Backprop and Kohonen in CommonLisp/Xlispstat. Both GPLed.7 80 .
http://www.geocities.com/researchtriangle/lab/5
Ask me about my kohonen-based image segmentation package done in Tcl/Tk.
Iuri Wickert
iuri.wickert@usa.net
Posted by Critical Mass:
o jekte/snns/snns.html
The Linux Journal has an article on the Stuttgart Neural Network Simulator. Great article! You can get the SNNS at http://www.informatik.uni-stuttgart.de/ipvr/bv/pr
There is a *great* Artificial Intelligence / Artificial Life HOWTO here geared towards Linux.
(I was only an egg, but then I cracked)
ThoughtTreasure understands questions and gives answers. Uses assortment of methods to analyze text, understand problem, apply common sense, and find answer to problem.
CLIPS rule and object expert system tool.
Knowledge Server Toolkit is a Perl-based system for monitoring and acting on continuous information flows, such as alerting when telemetry indicates unusual conditions.
If you're interested in feedforward neural networks with Bayesian learning. Radford Neal at the University of Toronto has a lot of nice code that runs on Linux/UNIX. The URL is http://www.cs.utoronto.ca/~radford
You can also check out Octave which may have somethings. This is a language similar to MATLAB. There is a toolbox that will work under MATLAB (and maybe Octave) from the Technical University of Denmark located here. By the structure of MATLAB and Octave, the source code is usually "open" (read: you can look at it) and it is in this case as well.
There is a plethora of open source neural networks out there, I myself wrote one in Scheme. SNNS is good so is Xnn/Nn (also does outputting of neural network c code). There is a FreeProp out there some place i think.
---------------------------------- I like fig newtons...they're tasty
I have just finished a Java implementation of a petri-net model for npc's in a game. It's not open source yet, but I think I can release the model by mid July
lroome@gulf.uvic.ca
I have poked around with jess. It's sort of based on CLIPS, and is programmable in lisp-like syntax. It also has hooks that allow you to extend it in Java or even call Java from within you Jess programs.
http://cslu.cse.ogi.edu/
The license for this project is not exactly "Open Source", but it is free for educational institutions. The neural net takes as input 5 LPC constants and has as output biphones (hundreds of them.) It's written in C. You can download it, play with it, and learn stuff. The license for commercial use requires payment.
I actually tried to make modifications to it in a class, but due to time constraints and lack of documentation I was unable to finish my changes. I think I made an A anyway and I learned a good deal about AI.
I spent many frustrating days recently downoading, compiling/porting and testing just about every freely-downloadable NN package I could get my hands on, for GNU/Linux, Win32 and Digital Unix -- just about everything that claimed to be a backprop simulator. Many were buggy or difficult to get up and running. Others were difficult to adapt or embed in a real application. Many didn't support a batch mode, or even worse, proper stopping conditions for training. Three clean, stable packages I found were: Aspirine/Migraines, Nevprop, and BPNN, in order of program complexity from most to least complex. However these programs had their limitations too, IMO. I just needed something robust, embeddable, reliable and fast for a University project.
The best software I found, which I cannot recommend highly enough, is MBP (Matrix BackPropagation) by Davide Anguista, available at ftp://ftp.esng.dibe.unige.it/neural/MBP/ . This uses highly-optimized matrix multiplication techniques for both forward and backward propagation, and the accompanying paper is worth a read just to see how the guy got his matrix multiplication so fast. MBP as a backprop simulator is 20-30 times faster than standard backprop NNs. And the code is clean, well-written, highly portable and stable. I take my hat off to Mr. Anguista.