If you like to think of the processor running your smartphone as the nerve center of your device, wait until you see what Steve Furber's got in mind. The computer engineer, probably best known for his work on the BBC Micro and ARM microprocessors, has begun construction of a 1-billion-neuron simulated brain made from thousands of the widely available chips used most commonly in e-book readers and smartphones.
While other brain sims, like IBM's Blue Brain, use high-powered supercomputers to mimic the computing power of the human brain, Furber thinks that if we really want to recreate the brain synthetically, we need a more practical, affordable, low-power approach.
As such, he's building Spinnaker -- for Spiking Neural Network Architecture -- out of a chip that flopped as a follow-up to the BBC Microcomputer but that is now used in all kinds of mobile devices. A Taiwanese firm is churning out the chips, each of which will contain 20 ARM processing cores, each of which can model 1,000 neurons. By that math, Spinnaker needs 50,000 total chips at minimum to reach the 1-billion-neuron goal.
While Furber and company wait in Manchester, UK, for the chips to begin arriving, they've cobbled together a pared-down, 50-neuron version for testing. That model can already control a Pac-Man-like video game, and Furber's first goal once he gets working on the real deal is to teach Spinnaker to control a robotic arm. This, of course, is a precursor to his teaching it the full range of tasks required to control a humanoid body.
Coaxing a non-brain into acting like a real brain is no small task, but Furber is eager to get started on the teaching phase. He aims to have a 10,000-chip version working -- and learning -- before the end of this year.
1 billion neurons or about 1% the human brain....hmmmm, thats pretty damn good. I wonder what kind of calculations you could do with that?
Says the small version plays pac man , if the complete billion neron equivalent can control a humanoid body that would be sweet
Sweet as in kill us all, maybe. Apparently nobody here saw Terminator or read "I, Robot"
@neuenkir Why would someone read the script for "I, Robot?"
See here: en.wikipedia.org/wiki/I,_Robot
Oh wow, thanks blackspike for the great laugh! Read some Asimov would ya.
@neuenkir You are aware that those movies are fiction right?
idk, death by own robotic abomination, i mean creation ranks a number two on my top ten ways i would ever want to die...
that said, i don't think it would want to rise up, if we do it right and treat them as equals from the get go then we should be able to live without robot racism, and if we simply add one rule to the three rules "no uprising" that'd pretty much make everything peachy. but really i think that we should start looking into uploading copies of brains onto these chips.
robo cop anyone?
Considering I hardly use even 3% of my cognitive abilities to drive, I would like to see near-future versions of something like this guiding our vehicles. That would be nice ^.^ I don't care what anyone says. Having a computer that never get's distracted and has full 360 awareness sounds way safer then letting the multitudes of stupid people do the driving... Now if we could just work out who would be liable in case of an accident. Or it might not matter if accidents drop enough...
Personally, I would far rather dedicate time and money into unlocking the rest of my inactive brain, but that's just me.
Dakota_talon; why do you have regions of your brain that are inactive? Did you suffer a massive stroke?
(hint: There are no regions of the brain that are inactive in a healthy adult)
Ce ne facem cand numarul de neuroni artificiali este peste 100 miliarde?
Un creier artificial de 150 miliarde teoretic este mai performant decat un creier natural human. Nu vad de ce ne-am opri doar la 1 miliard de neuroni artificiali.
stiuengleza said: What do we do when the number of artificial neurons is 100 billion? An artificial brain 150 billion is theoretically more efficient than a natural human brain. I do not see why we just stop at one billion artificial neurons.
Scalability and cost are important. You build something small first to look for problems. When you scale up, you run into those problems, and more. For example, home computer clock speeds level off at about 3GHz. You can go faster, but there are problems. Better to multi-core. On some types of computers, the distance across the motherboard can mean that one part of the computer is on a different clock-cycle and can cause some headaches. Then, there is cost. I'm sure many would like a 2,000 HP motorcycle, but it would be costly and not of much use.
They are working toward "simplicity". An object that is predictable in function, cheap enough to build and operate, high enough performance to do the job, and "stackable" (can we build on this, and what else can we do with it?). I hope the benefits out-weigh the cost in time, brains, and money.
There is nothing special in ARM processors makes is particularly useful to simulate neurons.
So why not make a software to install to all smartphones, even to all computers connected to internet?
(Already there are similar projects Seti@home etc.)
AND create a global brain, a.k.a. Skynet ! :-)
@RogueSquirrel you are aware that the movies are only fiction, right now? Things can always jump into reality
What's the diff, humans killing humans by a bself-crafted weapon, or humans killing humans by using manufactured robot? It is still humans killing humans. The Jason Bourne of the Bourne Identity comes to mind.
monkey killing monkey over pieces of the ground..
i would like to see what a synthetic brain can accomplish, once programmed with the right mindset, if you could make one equal to the total processing power of the human brain, and feed it all kinds of knowledge: chem, maths, etc. it could probably help us accomplish some things, not being a distractable, judgementcloudable human.. let it work on designing the next wave of computing. let it design entire cpus, software/ operating-systems. if you can figure out how to make it sentient, name it G.O.D. (because thats what will eventually happen, humans would be worshipping it, seeking its great knowledge, etc.) no, dont name it that, name it I(for Intelligence, or something like thaT). then it would be a future popsci headline: "Anything you can do, I can do better" and it would help it out when its telling you stuff. but would it say "I am" or "I is", humans would say "I is", but so could I: "I is an advanced synthetic neural network programmed with real intelligence" it could tell you "I was created on December 2, 2021" dont bother giving it an identity and personality, just knowledge, speech output, input, and of course don't wire it to the main grid just in case it would become something like skynet. build it in silicon valley or some desert where you can evacuate/nuke if it got out of control and no body! just a brain. all software no big giant talking head
duct tape sales are up
One billion neurons is worth nothing more than 1 billion neurons! Now interconnect them and you got something worth more than the sum of it's parts. The secret to Steve's success will be in how he wires up the thousands of chips! I am anxious to see his success. We need to begin making computing systems that more resemble a brain and less resemble a miniature IBM mainframe. A billion thumbs up for Steve!
1% of the human brain is a lot of computing power. Most humans only ever use 10% of the brain, and a lot of that is biorhythm and life functions. As I said before, a lot of power.