Original URL:
http://www.theregister.co.uk/2009/11/23/epfl_bluebrain_markram_modha/
IBM's cat-brain sim a 'scam,' says Swiss boffin
Neuroscientist hairs on end
By Austin Modine in San Francisco
Posted in HPC , 23rd November 2009 22:11 GMT
Responding to its nation's sovereign call to develop electronics with
the intelligence of a cat
(http://www.theregister.co.uk/2008/10/28/ibm_cat_brain_podule_scoop/),
IBM last week announced a major step in feline gray matter simulation.
But professional rivalries clearly run deep inside puss brain
replication circles, with a leading neuroscientist blasting the project
as a "scam" and a "mass deception of the public."
Dharmendra Modha, manager of IBM's Cognitive Computing unit, told the
Supercomputing Conference last Wednesday that he and his team created a
simulation of a cat-sized cerebral cortex
(http://www.theregister.co.uk/2009/11/18/ibm_closer_to_thinking_computer_chip/)
that included 1 billion neurons and 10 trillion individual learning
synapses. The project won itself the ACM Gordon Bell Prize in
recognition of outstanding achievement in high-performance computing at
the show.
But neuroscientist Henry Markram, the lead of the "Blue Brain" modeling
(http://www.theregister.co.uk/2005/06/07/brain_simulation/) project at
Switzerland's EPFL, has his claws out over IBM's claims.
"I am absolutely shocked at this announcement. Not because it is any
kind of technical feat, but because of the mass deception to the
public," Markam wrote in an open letter sent to IBM's technology chief,
Bernard Meyerson, and members of the media.
Markam claims the cat-brain simulation itself is relatively easy to do,
using only the "simplest possible equation you can image to simulate a
neuron." The results only demonstrate that IBM's Cognitive Computing
team has immense compute horsepower at their beck and call, he said.
All these kinds of simulations are trivial and have been around for
decades - simply called artificial neural network (ANN) simulations.
We even stooped to doing these kinds of simulations as bench mark
tests 4 years ago with 10's of millions of such points before we
bought the Blue Gene/L. If we (or anyone else) wanted to we could
easily do this for a billion "points", but we would certainly not
call it a cat-scale simulation. It is really no big deal to simulate
a billion points interacting if you have a big enough computer. The
only step here is that they have at their disposal a big computer.
For a grown up "researcher" to get excited because one can simulate
billions of points interacting is ludicrous.
He adds that "there is no qualified neuroscientist on the planet that
would agree that this is even close to a cat's brain" and that the
announcement is "simply a PR stunt here to ride on Blue Brain."
"That the Bell prize would be awarded for such nonsense is beyond
belief," Markam wrote. "I never realized that such trivial and unethical
behavior would actually be rewarded. I would have expected an ethics
committee to string this guy up by the toes."
IBM's Modha didn't get back to us directly when reached for comment, but
an IBM spokeswoman said the company stands by the statements that it
made regarding the announcement. She also notes that the brain
simulation wasn't solely done by IBM, but in collaboration with Lawrence
Berkeley National Laboratory and other research organizations.
"We made a cortical simulation at the scale that a cat brain functions.
We didn't say we made a cat brain that will go act like a cat," she
said. "That was never what we intended to do. This is all about looking
at new computing architectures - and looking at the brain to eventually
be able to create a computing architecture that has its low power
consumption and amazing computing ability. It's not artificial
intelligence we're after here."
Check out /IEEE Spectrum/ for a full copy of Markam's scathing letter
here