Slashdot Mirror
The Human Brain Project Receives Up To $1.34 Billion
New submitter TheRedWheelbarrow writes "The singularity looms as the Human Brain Project gets up to $1.34 billion in funding. 'The challenge in AI is to design algorithms that can produce intelligent behavior and to use them to build intelligent machines. It doesn't matter whether the algorithms are biologically realistic — what matters is that they work — the behavior they produce. In the HBP, we're doing something completely different...we will base the technology on what we actually know about the brain and its circuitry.'"
Yeah.
In the HBP, we're doing something completely different...we will base the technology on what we actually know about the brain and its circuitry.'"
With this approach, they will probably start with nematode brains.
And realize they don't have to go any farther.
Faster! Faster! Faster would be better!
Why study a human brain?
The more ways we attack a given problem, the more chances of success. We have different communities working on different approaches to AI: Statistic, symbolic and biologically inspired. All three have produced interesting results already, meaning they have solved some practical problems.
Also, most human brains can show "intelligent behavior" in certain ways that our latest algorithms can't, e.g. navigating an arbitrary kitchen and finding a beer in the fridge :-)
When his defense asked, "Which computer has Jon Johansen trespassed upon?" the answer was: "His own."
I believe that what they receive is actually up to 0.5 B€ in matching funds, meaning that for every 1 € they get from other sources (private persons, foundations, national funding bodies, etc...), they will get another 1 € from the EU, up to 0.5 B€ for a total of about 1 B€. Also this is granted under the EU Framework Program 7 which ends soon. So really what they got so far is 54 M€ for 30 months and the rest will come after that under the new EU program/package (Horizon 2020) which is currently being negotiated. Given the financial health of EU countries right now, there is a chance that the overall envelope is cut down and it is not clear how much funds they will get from national bodies in the first place.
The EU is also funding under the same initiative another B€ project about graphene.
The Human Brain Project promises a lot (AI, curing neurodegenerative diseases, understanding the brain and consciousness, limiting animal experimentation, etc...) and it is the opinion of most neuroscientists in the US and in Europe that it won't deliver. If you google it, you will find many interviews from neuroscientists who are very critical of it. It is difficult to evaluate what really will come out of it.
Massive large projects like this almost always end in utter failure. Even the IBM cat brain project failed to accomplish much. Intelligence is much more complicated than a mere randomly connected neural network. I just hope something good comes from this and it is not a total waste.
But the following proviso is misguided: "It doesn't matter whether the algorithms are biologically realistic--what matters is that they work--the behavior they produce."
The basic algorithm to produce human behavior is essentially biological:
10: Wine
20: Women
30: Song
40: GOTO 10
Sex, drugs and rock & roll for you hipsters out there (and quit trespassing on my lawn to collect magic mushooms).
Set your phasers on "funky"!
I absolutely am in favor of basic science research, but looking through their documents, I can't find the answer to this problem.
What is the success metric? They have a system, which is basically a super computer, and they will have it solving some equations. The equations represent some parts of neurons, but not all. How will they know that they've succeeded? The computer isn't going to simulate any real human brain, we don't know what that looks like. We barely know what C. Elegans' looks like. Are they going to use this computer to answer some question? What question?
What are they going to use to know if they've succeeded? Overly-optimistic promises are what killed a lot of AI research around the 1970s.
"First they came for the slanderers and i said nothing."
Human brains may be weak, but the vision recognition algorithms are amazing.
"First they came for the slanderers and i said nothing."
It seems unclear to me that human brains produce "intelligent behavior." It seems to depend on the brain. Only a few per hundred seem to work really well, but up to half of them can file TPS reports.
The popularity of TV shows like "Here Comes Honey Boo Boo" and "The Housewives of _______", not to mention the people actually *on* those shows, would seem to support your thesis.
It must have been something you assimilated. . . .