billnet - Summary
This group is not part of the GNU Project.
Billnet is a modular neural network simulator platform. Its primary goal is to simulate as many as possible of the existing architectures and algorithms for neural networks in an effective and light weight manner so billnet could be used simultaneously for research, educational and industrial solutions.
Registration Date: Sun 30 Mar 2003 09:03:34 PM UTC
License: GNU General Public License v2 or later
Development Status: 0 - Undefined
posted by vasvir, Wed 05 Jan 2005 04:49:29 PM UTC
A lot of work is underway to port old billnet functionality to the newer codebase. KNN, BFGS*, Conjugate Gradient*, Levenberg-Marquardt are already ported. FLF is splited to a library for the most part and the part continues for billnet 0.7
.Bill
- we have some numerical discrepancies here and there...
posted by vasvir, Wed 09 Apr 2003 09:46:59 PM UTC
ALECO (0 and 2) by S. J. Perantonis and D. Karras are the algorithms that pushed the first order gradient algorithms to the limit.
posted by vasvir, Sun 06 Apr 2003 09:05:51 AM UTC
The back propagation family is the base of all gradient learning machines. The current billnet's implementation can be used with:
-the cross entropy cost function
-the online and offline type of learning
-the sigmoid output function
-the tanh output function (untested)
posted by vasvir, Sat 05 Apr 2003 07:14:22 PM UTC
We are slowly getting our scripts up to speed with savannah infrastructure. Hopefully more algorithms will follow soon.
- Browse open items
Submit a new item
- Browse open items
Submit a new item
Powered by Savane 3.14-60ba.
Corresponding source code