Warehouse D
Artificial Intelligence
Page 2
Page 3

Egad - They Simulated A Mouse Brain On A Computer

By James Donahue

It sounded relatively simple when the news broke, but what three researchers at the IBM Almaden Research Lab and the University of Nevada recently accomplished is quite astounding. Using a supercomputer they simulated the thought patterns of half a virtual mouse brain.

If science can do this, is it possible that a computer may someday exist with the thinking capacity equal, or superior to that of the human brain? Will we someday build robots that can not only live with us, but possibly out think us?

Yet a third scenario comes to mind . . . will it be possible for the human spirit to transfer from these delicate and perishable bodies into robots designed to exist on an overheated and polluted world, and even endure long trips through space to other stars? Such are the dreams of such thinkers as noted physicist Stephen Hawking.

The road to building a computer that simulates the human brain is going to be a long one, but the fact that researchers are taking it is good news, indeed.

The three people working on this project are James Frye, Rajagopal Ananthanarayanan and Dharmendra S. Modha.

In a brief paper titled "Towards Real-Time Mouse-Scale Cortiacal Simulations," they said they ran the simulation on a BlueGene L supercomputer that used 4,096 processors, each one of them using 256 megabites of memory. In this experiment, the researchers said they created half a virtual mouse brain that had 8,000 neurons with up to 6,300 synapses.

By comparison, a healthy and fully active adult brain has an estimated 100 billion neurons operating.

Even with this limited experiment, the simulation was so complex that it was only run for ten seconds and at a speed ten times slower than a mouse brain operates in real life. Thus what was achieved was the equivalent of one second in time in the brain of a mouse.

Thus the work of constructing a computer that mimics the electrical impulses of a human brain has a long way to go before it is actually achieved. Yet there is an urgency involved that should be driving this research. It also is troublesome to think that possibly this single team of three people is the only one attempting to achieve this goal.

For those who might worry about the day when the robots take over the world . . . the threat is not upon us just yet.