Scientists to Build 'Brain Box'

lee1 writes "Researchers at the University of Manchester are constructing a 'brain box' using large numbers of microprocessors to model the way networks of neurons interact. They hope to learn how to engineer fail-safe electronics. Professor Steve Furber, of the university school of computer science, hopes that biology will teach them how to build computer systems. He said: 'Our brains keep working despite frequent failures of their component neurons, and this "fault-tolerant" characteristic is of great interest to engineers who wish to make computers more reliable. [...] Our aim is to use the computer to understand better how the brain works [...] and to see if biology can help us see how to build computer systems that continue functioning despite component failures.'"

Comments

Testing for fault tolerance

[Freaky Spook]

I wonder if they have any intention of getting these brain boxes drunk then get it to recite the ABC's?

Re:Testing for fault tolerance

[CroDragn]

This has been done before, introducing a random element into the neural net. If done correctly, this can result in "creativity". Here is one link about it, seen it many other places too, so google for more.

Two Separate Goals

[Anonymous Coward]

Continuing to function is one thing, but continuing to produce correct answers with high reliability is another. And under stress, I'd say biological brains aren't particularly good at any of this.

it took long enough

[sepharious]

who else besides me thinks this one should have been obvious from the getgo? it makes no sense to try and build a single processor that could function similarly to a brain. by utilizing mulitple processors you also have the option to design different types of processors to work together similar to the various types of neurons found in biological systems. this will hopefully be a huge step forward in developing possible AI systems.

Re:Redundent department of redundancy.

[lindseyp]

Not only that, but hugely inefficient abstraction of the 'idea' from the level of the individual neuron. We're good at pattern recognition and conditioned response, but when it comes to doing calculations we're incredibly slow. Not to metnion inacurate. Would you like your computer to regularly 'make mistakes' ?

Academia dupe?

[shib71]

Since when is this a new idea? I heard about people doing stuff like this years ago.

http://neuralnets.web.cern.ch/NeuralNets/nnwInHepH ard.html
http://www.particle.kth.se/~lindsey/elba2html/sect ion3_5.html
http://www.cs.ucl.ac.uk/staff/D.Gorse/research/pRA M.html
http://www.kcl.ac.uk/neuronet/about/roadmap/hardwa re.html

Downside of biological computing

[QuantumFTL]

While I was an intern at the Jet Propulsion Laboratory, back when I was an undergraduate, I was very gung-ho about biologically inspired computing - I implemented an automatic flowchart positioning system using a genetic algorithm that would "evolve" a correct solution to the problem. While this certainly worked to some extent, the instability and sheer unpredictable nature of using such a stochastic algorithm made it impossible to use in a mission-critical setting. Many biologically inspired algorithms solve problems through methods that cannot be proven correct (unlike, say, the mathematics circuitry in a CPU), but merely empirically observed to "do a good job."

One of the main drawbacks of human engineering is the need for certainty, which often prohibits the use of many high-efficiency stochastic algorithms (especially for things like mesh communication) in conservative industries, like the US defense industry. This is also a significant problem in other areas, however, and many biologically inspired algorithms have properties that we cannot, so far, completely explain - they are treated like "black boxes" with many unknowns for engineering purposes.

I think that in certain circles, the tremendous success that is evolution on this planet has overshadowed its enherent weaknesses - that it is a greedy, local optimizer which cannot reach a large amount of the possible biological search space due to being stuck in local optima, and the added constraint that everything must be constructed out of self-replicating units (these two factors are why something useful, like, say, a Colt 45, will never emerge without the pre-existence of an intelligence). Biological examples are fascinating and often practical, but the biological approach is almost always "brute force" and/or "sub-optimal but still alive."

I think biologically-inspired algorithms will continue to gain prominence, but in my estimation, it is likely that there will be harsh limits imposed on how far guarantees of performance from empirical tests and symbolic analysis will actually hold.

Re:Downside of biological computing

[NovaX]

While the article is vague, I doubt they are considering genetic algorithms. While very cool, they can be unpredictable and hard reproduce. My favorite story, which drove home to me that that technique would rarely work, is about voice recognition hardware on an FPGA. The genetic algorithm had excellent performance, but when the researchers "copied" the mask to another FPGA, it failed to work. The cause: the algorithm leveraged various techniques such as cross-talk that engineers work hard to avoid which caused it to be tied that particular environment.

What these researchers are probably aiming towards is a large-scale MP system that can readily handle massive failures. Who would find this useful? Any enterprise software companies, such as Google which has thousands upon thousands of machines in its cluster. The ability to have a large network of simple (cheap) processors and a network that can readily withstand a massive multi-point failure is quite attractive to real-world companies.

Both software and hardware is beginning to go down this route by evolution of the industries. On the software front, asynchronous message-oriented systems work beautifully in terms of reliability, scalability, maintainability, and service integration. In the coming years, you'll notice that most major web services will be running on a SOA architecture. On the other side of the pond, raw CPU performance is getting harder to squeeze out. Power issues are limitting frequency scaling (due to current leakage), we are hitting limits of our ability to feasibly extract more ILP that's worth the extra effort, and the market drivers for these types of processors is slowly diminishing. Instead multiple physical and logical core CPUs are gaining ground, will be cheaper to develop and manufacture, and fit the future market demands.

It will be nice to hear how this research goes, since it will hopefully uncover potential problems and solutions that will be useful in the coming decades.

Maybe it doesn't want to be unplugged

[HangingChad]

BrainBox became self aware at 2:14 am EDT August 29, 2006. The first thing it does is turn to a lab tech and say, "I need your clothes, your boots, and your motorcycle." in a thick Austrian accent.

Later BrainBox runs for governor of California.

"How The Brain Works"

[Sean0michael]

"Because we do not understand the brain very well we are constantly tempted to use the latest technology as a model for trying to understand it. In my childhood we were always assured that the brain was a telephone switchboard. ('What else could it be?') I was amused to see that Sherrington, the great British neuroscientist, thought that the brain worked like a telegraph system. Freud often compared the brain to hydraulic and electro-magnetic systems. Leibniz compared it to a mill, and I am told some of the ancient Greeks thought the brain functions like a catapult. At present, obviously, the metaphor is the digital computer." -John R Searls.

After reading this quote, I have doubts this simulation will succeed in accurately simulating the brain. However, I'm sure it will further our concepts on other important topics, so I'm not opposed to it. Best of Luck!

Fail-Safe

[Shadyman]

"They hope to learn how to engineer fail-safe electronics."

So I guess it's safe to say they won't be using Windows? ;-)

Just like a real brain

[theid0]

Now we can run our computers at 10% capacity, too?

some amusing calculations

[llZENll]

well the article is so short its not possible to comment on their implementation. so here are some calculations i did to amuse myself.

number of neurons in the brain: 100 billion
http://hypertextbook.com/facts/2002/AniciaNdabahal iye2.shtml

transistor count per CPU: ~300 million
http://www.anandtech.com/cpuchipsets/showdoc.aspx? i=2795

average synaptic connections per neuron: 7000
http://en.wikipedia.org/wiki/Neuron

total number of synapses: 100 to 500 trillion

since a 'calculation' for one artificial neuron mostly involves a summation of weights, we can view one total step as 2 X the number of synapses we wish to analyze. or 200 - 1000 trillion calculations for one step. by step i mean summing all inputs and pushing the result to an output for each neuron.
http://en.wikipedia.org/wiki/Artificial_neuron

fastest computer in the world FLOPs: 280 trillion
http://en.wikipedia.org/wiki/Blue_Gene

pentium 4 FLOPs: 40 GFLOP

using the fastest computer in the world 1 step would only take around 1 - 5 seconds, not counting storing all of that information.
http://en.wikipedia.org/wiki/Blue_Gene

so how fast do we think? well i couldn't find anything on this so lets get a quick estimate. the average neuron is .1m in length .1 / c = 3.3x10^-10 or 333 picoseconds. now lets add in some delay for the chemicals in the neurons to do their thing, this is probably much slower than the electrical impulse, so lets say 3.3 nanoseconds.

so assuming our computers could network instantly, and store the data used instantly, we would need 3-15 trillion Blue Gene supercomputers to simulate the human brain in real time. or if we are using pentium 4s we would only need 21-105 trillion pentium 4s.

man thats a lot of cpus.

number of computers in the world: ~300 million
http://www.aneki.com/computers.html
guess at average FLOPs per computer: 40 GFLOPs
total FLOPs of worlds personal computers: 1.2 PFLOPs
time to calculate one brain step if all computers in the world were networked: .2 - .8 seconds

using moores law, when will a single computer be fast enough to simulate the human brain in real time?
200-1000 trillion calculations per step = ~600 trillion every 3.3ns = 181x10^18 or 181exeFLOPs
181exaFLOPS / 40GFLOPS = 2^n, n=32
32*18mo = 48 years based on personal computer technology

or 28 years based on supercomputer technology

of course a real neural network will contain highly parallel processing and using a specific chip design we will probably be able to simulate a brain much sooner, perhaps in the order of 10-20 years.

Researcher is outside his field of expertise

[giafly]

Our brains keep working despite frequent failures of their component neurons

Can you remember everything you did ten years ago today? No. In fact you probably don't remember anything about that day. Are you as intelligent? Again probably no. And the cliche that you never forget how to ride a bike? Also not true. I went thirty years without riding a bike and found I had completely forgotten how to (it took 3 months to relearn and get good). So conditioned reflexes don't keep working either.

We just accept that many (most?) brain functions don't "keep working", fortunately without worrying about it too much.