11 comments

[ 3.5 ms ] story [ 33.7 ms ] thread
If this is true that it is simply an Artificial Neural Network, then it really is a scam. I was under the assumption that it was part of the IBM Blue Brain project, but it is obviously not. For some real ground-breaking research, take a look at Blue Brain: http://bluebrain.epfl.ch/
(Disclaimer: I work for the Blue Brain Project. What I say is my own opinion.)

It's not just an ANN like you may have learned in school. They refer to their model as "single-compartment phenomenological spiking neurons, learning synapses with spike-timing dependent plasticity, and axonal delays" in their The cat is out of the bag article.

In simpler language, it's a spike-based model (when someone says "artificial neural network" they usually mean a rate-based model) where they adjust synaptic weights based on the correlations of spikes (Google: spike-timing dependent plasticity), and add a transmission delay before a spike arrives at the post-synaptic cell.

It sounds very much like the model described in Hebbian learning and spiking neurons (Gerstner et al., 1999).

This is a more sophisticated model than most neural network models, but it's still very much abstracted from biology. For starters, it doesn't sound like they model a diversity of neuron structures, nor even ion channels or dendrites.

This is where this "Blue Matter" project differs from the Blue Brain Project, which is attempting to model the biology is as fine detail as the science (and the computing hardware!) allows.

EDIT (forgot to get to the point):

Is it a scam? That's open to interpretation, but I agree with Dr. Markram that they are making fantastic claims and glossing over the fact that they are not modelling anything resembling a cat's brain except in neuron count.

You work for the Blue Brain project? That's amazing. Do you have a blog or anything where I can find out more? Can you tell us more about what it's like? I think it's one of the most fascinating projects going on right now.
I'm just finishing up an internship with the project, and I don't keep a blog. Sorry :)

That said, it's a very cool project with a great group of smart people. I'm very optimistic about its future.

I follow AI quite a lot and I'm curious if there is much talk about the future of memristor and brain emulation within the project seeing as a hardware variable analog would seem the ideal medium for such things. I know when this technology came out AI was what first came to mind.

Working for the Blue Brain Project would be my dream job as a programmer/engineer. I'm curious how you came about your internship? I'm desperately looking for a challenge and this type is research would definably be my first choice.

That would be my dream job too. One obstacle is that it's in some foreign country, Switzerland I think?
Looks like the source code to this already leaked:

1 PRINT "Meow"

2 GOTO 1

This is awkward. I hope someone investigates this thoroughly.
"I think the proper response here is: mee-yow!"

No, the proper journalistic response is to investigate the facts and find out what is really going on.