Connectionists: bnns 2.0.beta4 (Research tool for interactive training of artificial neural networks)

[matus.uzak@gmail.com]

Hi,

I would like to announce bnns 2.0.beta4 release, [1].

bnns is a research tool for Interactive Training of Artificial Neural
Networks based on the Response Function Plots visualization method. It
enables users to simulate, visualize and interact in the learning
process of a Multi-Layer Perceptron on tasks which have a 2D
character. Tasks like the famous two-spirals task or classification of
satellite image data.

Most noteworthy features include:
- reset, freeze and unfreeze of weight coming into a neuron
- visual inspection of extrapolation performance
- a set of self documented Perl scripts which assist in patterns
preparation process
- a Perl script to prepare bnns patterns from Boston Remote Sensing
Testbed [2] data

But there are many other features which enable visual analysis of NN
performance, simply take a look at the manual draft.

The main purpose of the project is to help students to understand the
knowledge representation in ANNs and the process of it's formation
with different setup of training patterns, training algorithm
parameters and NN topology.

And there's still time to address any comments into the project
mailing list [3] to be included into the 2.0 release.

Enjoy!

[1] https://sourceforge.net/projects/bnns/
[2] http://techlab.bu.edu/resources/data_view/boston_remote_sensing_testbed/
[3] bnns-devel at lists.sourceforge.net

br,

Matus Uzak