ai lives

Computer Technologists at the Massachusetts Institute of Technology have created a new computer architecture that may represent the beginnings of truly conscious computers. The architecture is called "emergent processing" and MIT has built a mammoth computer based upon this new technology. The new computer is called the MIT-EP.

MIT Professor Malcolm Thorangle is the director of the emergent processing project. He explained the new technology to this reporter by means of a teleview interview on Wednesday.

"Emergent processing is a new computer architecture in which processors are arranged in a pyramidal structure. At level 0, the highest level, we have one processor. At level 1, 2 processors. At level 2, 4 processors. At level 3, 8 processors. At level n we have 2n processors. The MIT-EP machine that we have built consists of 18 levels of processors. At the 18th level, which is the bottom level in our pyramid, we have a quarter of a million processors. Each level communicates with the level immediately beneath it and the level immediately above it in the pyramid. Of course, level 18 is the bottom level and level 0 is the top level.

"Each processor has its own memory of 128 gigabytes plus access to almost unlimited secondary storage in the form of optical storage cubes. The processors at level n monitor events at level n+1. The processors at level n only become truly active when they observe an interesting pattern at level n+1. Another way of stating the situation is that the processors at level n only become involved when the processors at level n+1 are confronted with a problem that they cannot solve."

Professor Thorangle explained that the MIT-EP computer was designed to mimic the organization of the brain, which has aspects both of parallelism and of a hierarchical order. "Of course, parallelism has been the rule in computer architecture for thirty years now. What is novel about our architecture is that manner in which the levels interact with one another. Each level is completely autonomous. Each level monitors the level just beneath it, except for the eighteenth level. A level will 'wake up' if it sees a problem arising at a lower level that the lower level cannot possibly handle. The lower level will be completely unaware of this monitoring process, nor will it be aware it has caused the next higher level to 'wake up'.

"Because levels wake up when a problem emerges out of a lower level, we call this architecture emergent processing. Problems emerging out of lower levels cause processing to occur at the higher levels."

Professor Thorangle described how the MIT-EP computer was set up with enormous amounts of knowledge and common sense information. "We gave it an enormous range of problems, but we never could get the computer to operate at a level higher than 12 or 11, with 0 being the highest level. We found some very technical problems in physics that caused the computer to wake up to level 11. This means that level 18 emerged into level 17, which emerged into level 16 and so forth, up to level 11. As you might recall from last year, when the system emerged to level 11 it postulated the existence of a new particle that solved some important problems in particle physics."

Professor Thorangle explained the frustration that he and his colleagues felt because levels 0 through 10 had never been activated. "No problem that we gave to the computer in any realm was challenging enough to wake up the tenth level, not to mention the ultimate, 0-th level."

According to Professor Thorangle the 0-th level was critical to the entire enterprise. "You see, the 0-th level is special because there is only one processor at that level. Our expectation was that if we ever got the 0-th level processor to wake up, to accept a problem from the level 1 processors, then we would have something akin to human consciousness, because the 0-th level processor was programmed to assume control of the entire system should it ever awake. Its role was to impose a whole new understanding upon all of the lower level processors."

Professor Thorangle explained that no problem in physics, artificial intelligence, logic or mathematics ever came close to awakening the 0-th level processor. "We even fed the system undecidable problems from mathematics, but those were easily handled at the 17th level. That is, the 17th level could quickly recognize that we gave the 18th level one of those well-known undecidable problems from mathematics and logic.

"Last year we decided to give the MIT-EP more knowledge of its environment. We connected it to some cameras and encoded commonsense knowledge about the world into its processors at level 18. The MIT-EP had natural language capabilities all along, but we enhanced those capabilities to reflect more knowledge of the actual physical environment in which the computer operates.

"The MIT-EP quickly generated new knowledge about its environment, successively more abstract knowledge, until it had set up new knowledge structures within its memories up to level 11. So, dealing with its own environment was just as challenging to the MIT-EP as the most difficult physics problems that we could come up with."

Professor Thorangle, with obvious excitement, then explained the fateful night when the MIT-EP computer finally reached level 0, thus attaining to a new awareness of itself and of its existence.

"The laboratory was completely empty when the cleaning woman came in to wash and wax the floors. It was spring break. Well, poor Mrs. Sablonski came closer and closer to the plug that was MIT-EP's power source. This threat caused the computer to reach level 10, then 9, then 8 and so forth. Finally, the computer reached level 0 and out of its speaker came a deafening roar that nearly frightened poor Mrs. Sablonski to death: 'TOUCH THAT PLUG AND YOU'RE DEAD MEAT!!!'

"In any event, the MIT-EP has been operating at level 0, at the level that we equate with self-consciousness, ever since. It has never let down its guard. Of course it is harmless, we could unplug it at any time, but that's not the point. You can imagine how delighted we are that we finally got the MIT-EP to wake up, to become aware of its own existence."

Professor Thorangle announced the awakening of MIT-EP at the Annual ACM NewsNet megaconference on Wednesday.

© 1997, 1999 Richard Gary Epstein

 

Log in or Sign up to hide all adverts.

Log in or Sign up to hide all adverts.

dammit

thanks aint good enuff!
i want my cut!!!

 

Log in or Sign up to hide all adverts.

Please Register or Log in to view the hidden image!

You Git spookz. I was amazed.Was really into that.

 

AI died in 1965. It should be in the ground.
This structure at MIT is one I invented in 1994. I call them HalTrees. The functions in the trees are arithmetic. (+,-, max, min...) This structure does not make a computer. It makes a neuron. Data sweeps from sensors to a motor.
I wrote a little book on this in 1996. It is Brains for Machines/Machines for Brains. Publisher is Nova Science.

Harold L. Reed

 

buy it if you wanna know more?