X-Message-Number: 21408 Date: Sat, 15 Mar 2003 19:26:51 -0500 From: Francois <> Subject: About brains and computers I would like to start by saying that brains and computers are not the same type of system. Computers process information with impulses that are combined according the the rules of bolean logic. Those impulses are generally electrical in nature, although they can take other forms, such as moving gears and levers, fluid flows, interfering photons, etc. Electric impulses are merely our most developped technology. Brains are communities of specialized cells, chiefly neurons and glial cells, which are interconnected in complex ways and which also maintain around them the biochemical environments they need to function adequately. It seems that the rate at wich these cells fire their impulsions, as well as their timings, are the chief features enabling them to process information. There are no 0's and 1's in brains. Brains are material objects that strictly obey the laws of physics. Provided we understood the basic principles by wich they work, we could make machines who worked with the same principles and therefore create true machine supported intelligence, self awareness and feelings. The question is, can we achieve this with computers? Well, there is nothing preventing us from programing a digital model of a neuron and puting it in a digital representation of a brain's biochemical environment. We can even have such a model translated as an image on a monitor screen and "watch" the "neuron" recieve impulses and respond by firing its own impulses. Provided we have fast enough computers with enough memory, we could interconnect billions of these neurons and build a simulated brain that way, even constructing retinas, olfactory bulbs, inner ears and all the arrays of specialized neurons we use to gather information from our environment, all digitally represented in the computer's memory. If we then compared the workings of this simulated brain with the workings of a real one, we would see them behaving in much the same way. For instance, nerve impulses triggered by light hitting a real retina travel in the various layers of our brains, where they are correlated, interpreted and translated as visual information. We would witness the same thing happening to the simulated impulses triggered in the simulated retina by simulated photons. Our digital representation of a brain would produce outputs similar to those of a real brain if provided with similar inputs. I think such a digital model would be just as self aware as we are, even though it would be only a collection of bits being manipulated by an advanced digital processor. We are still many years away from having the computing power needed to build such a model, and still more years away from having the understanding required for the task. But we are fast getting there. However, I don't think it will ever be a practical way of making a sentient machine. I believe the analog path will be better than the digital one in this case. Francois --------------------------------------------------------------------- No lifespan shorter than eternity is acceptable --------------------------------------------------------------------- Rate This Message: http://www.cryonet.org/cgi-bin/rate.cgi?msg=21408