[ml] pybrain and deep nets next?

[Mike Schachter]

(Carrying on this conversation from previous threads)

Unless anyone objects, it seems like deep nets are the
way to go...

Next week I figure a good thing to do is to have an
installfest and start experimenting with PyBrain:

http://pybrain.org

They claim to have implementation of deep belief nets and
restricted boltzmann machines. We should work off of their
latest source code, accessible via git:

git clone git://github.com/pybrain/pybrain.git

An alleged example using RBM's can be found in
in examples/unsupervised/rbm.py, which we should
all examine at some point.

I just joined their mailing list, and will probably be asking
them questions about the implementation. Their mailing
list is accessible through their google groups page:

http://groups.google.com/group/pybrain

So, see everybody Wednesday!

  mike