Tuesday, 11 August 2009

Artificial Brains Within 10 Years

As I've blogged about before, there are two approaches to AI - analytic and synthetic.

Here's some news on the progress made in the Analytic approach, where instead of trying to "grow" and AI, we analyse what happens at the cellular level of existing intelligences, and try building a brain out of simulated cells. From RedOrbit :
A leading scientist has claimed that a detailed, functional artificial human brain could be built within the next 10 years, BBC News reported.

Henry Markram, director of the Blue Brain Project, has already simulated elements of a rat brain.

A synthetic human brain would be of particular use finding treatments for mental illnesses, Markram told the TED Global conference in Oxford.

He said some two billion people are thought to suffer some kind of brain impairment and that it is not impossible to build a human brain within 10 years time.

Launched in 2005, the Blue Brain project aims to reverse engineer the mammalian brain from laboratory data, particularly focusing on the neocortical column – the repetitive units of the mammalian brain known as the neocortex.
He likened it to cataloguing a bit of the rainforest, as in how may trees does it have, what shape are the trees, how many of each type of tree do we have, what is the position of the trees.

"But it is a bit more than cataloguing because you have to describe and discover all the rules of communication, the rules of connectivity," he added.

His team has been able to digitally construct an artificial neocortical column with a software model of "tens of thousands" of neurons - each one of which is different.

They have found that the patterns of circuitry in different brains have common patterns even though each neuron is unique.

Markram said we do actually share the same fabric even though each brain may be smaller, bigger, or may have different morphologies of neurons.

"And we think this is species specific, which could explain why we can't communicate across species.

The team feeds the models and a few algorithms into a supercomputer to make the model come alive and Markram said since you need one laptop to do all the calculations for one neuron, they would need ten thousand laptops.

But instead they use an IBM Blue Gene machine with 10,000 processors and the simulations have started to give the researchers clues about how the brain works.

For example, they can show the brain a picture of a flower and then follow the electrical activity in the machine, where it creates its own representation.

He said the ultimate goal is to extract that representation and project it so that researchers could see directly how a brain perceives the world.
One thing - the Human brain cas about 100 Billion neurons. However, today's average desktops can simulate maybe a dozen each.

I don't know whether to be amazed at how complex the human brain is, or how simple. Because we'll soon have 10 Billion people on the planet. And if each one had an average desktop of the time, each communication with the 20 or so in the local area... yes, there would exist a network capable of modelling at a fine-scale a human brain. With about the same reaction times too.

A Brain the size of a planet.

I wonder...

More in this at the Blue Brain project. How appropriate.

No comments: