Brainy Magnus sees a few of the options

October 31, 1997

FOR Igor Aleksander of Imperial College, London, artificial intelligence is only really of interest if it tells us something about human intelligence.

Software, he says, can recognise patterns, such as faces, financial data, and medical and geophysical images, and it is possible to make a machine that gives Gary Kasparov a chess game he will not necessarily win.

But none of this is that exciting, he believes. "Research has produced a lot of useful know-ledge and products. But artificial intelligence has now lost its identity and includes a lot of material not related to its core, which is the creation of programmes whose activity would be said to require intelligence if humans did them."

Professor Aleksander says, however, that new methods in neurological science are opening doors for AI, allowing us to see what the brain does, and try and imitate it in computers.

His preferred brain function is the vision system. "It is a very rich system and a lot is known about it," he says.

He uses data on how the brain responds to visual inputs as the template for Magnus, a computer which uses about one million electronic "neurons" (the brain has 10,000 times as many) and about 100 inputs to each, compared with perhaps 6,000 to 7,000 for a human brain neuron.

There is no worry about the difference in scale between Magnus and real brains, since "the mathematics says that the scaling between the two is fine". More to the point: "It keeps us in awe of what the real brain does."

Yet even at this level of simplification, he says, the development of something like free will appears, with some unpredictable decisions emerging when Magnus is offered a choice of action.

If offered food or drink while thirsty, Magnus will always go for drink, but if it is neither thirsty nor hungry, the choice becomes unpredictable.

Using machines like Magnus, says Aleksander, allows us to bypass "messy" problems such as those linked to philosophy. For him, concepts like awareness or consciousness, are just things the brain does, and the mind/body split that fills journals is meaningless. "I am an engineer, and as I see it, everything we see can be ex-plained by neural activity," he says.

Research papers

Research papers relating to the main stories on this page can be found on the following sites:

http://www.dai.ed.ac.uk/ (Edinburgh)

http://www.cogs.susx.ac.uk/users/inmanh/ (Sussex)

http://www.ee.ic.ac.uk/research/neural/neural.html (Imperial)

You've reached your article limit.

Register to continue

Registration is free and only takes a moment. Once registered you can read a total of 3 articles each month, plus:

  • Sign up for the editor's highlights
  • Receive World University Rankings news first
  • Get job alerts, shortlist jobs and save job searches
  • Participate in reader discussions and post comments
Register

Have your say

Log in or register to post comments