4.7.2011

STEPHEN FURBER’S CURRENT PROJECT IS UNEARTHING BRAIN SECRETS

Professor Stephen Furber, the Millennium Youth Camp’s guest of honour, introduced MY Campers to the wonders of brain simulation. Even though partial simulations of human brain functions are already feasible, the scientific community is still far from being able to construct a complete artificial brain.

The aim of the SpiNNaker project, Professor Furber’s current undertaking, is the building of a network that simulates the functions of the human brain. Using interconnected multi-core processors, he and his fellow researchers will soon be able to build a system capable of parallel information processing in real time.

“The brain’s capabilities and its plasticity in the parallel processing of information represent both core aspects of learning and the roots of the brain’s efficiency,” said Professor Furber. The brain is also superior to information technology in other respects. ”In addition to its energy efficiency, the brain is a more sophisticated and more functional system than anything man has ever produced.”

Is it time for a biological computer?

The SpiNNaker project is based on the assumption, widely agreed among neuroscientists, that the human brain can be viewed as a constantly-modulating network. Even so, the brain’s basic building blocks are neurons, the cells that pass information on. When MY Camper Sagar Khavnekar asked the Millennium Prize Laureate for his opinion about the idea that biological computers are the future of information technology, Professor Furber slated the idea as unachievable.

”Biology-based computers cannot handle complex issues. And they are too slow,” he said. On the other hand, he conceded that one feature which biology can be used to achieve is massive parallelism. ”Unfortunately, parallelism is one of the things we have no idea how to use,” he continued.

Just one human brain requires several nuclear reactors

If SpiNNaker fulfils its potential, it will be able to simulate a network consisting of a billion neurons. And even though this equates to just 1% of the neurons in the human brain, running it requires a lot of processing power and therefore energy.

“To simulate the functions of a complete brain using current high-performance computer technology, the power output of several nuclear reactors would be required,” said Furber. While both computer performance and the levels of energy efficiency achieved have developed at a remarkable pace, they are still a very long way from offering a practical replacement for the human brain.

And although the individual components which make up our brains may be slow, the aggregate is pretty efficient. Even a partial brain lesion does not result in a significant reduction in speed. ”Appreciate your brain,” was Professor Furber’s advice to MY Campers.