Modern robotics developed concurrently with cybernetics, as well as information theory
and computer science. In the sixties and seventies, certain hybrids
began to emerge, like artificial intelligence (A.I.) and neural networks. There is a big difference between artificial intelligence and neural networks. As Ray Kurzweil explained it, A.I. moves very sluggishly because it tries to map out every possibility for a robot’s movement in a kind of logic tree through trial-and-error, or heuristics. There was much promise for A.I. in the seventies and eighties, but now things have cooled.
Neural networks is apparently the hot topic now, which is more computer science based. Kurzweil predicted that when the grunt-work of the top-down, hierarchical A.I. methods met the bottom-up approach of neural networks, robot sophistication will increase exponentially .
At least for the next half-century we’re safe, apparently. Whew. Until then, robots perform surgery, filtering out the natural shaking of a surgeon’s hands, explore volcanoes, distant worlds, and the Antarctic. Some try to pitch themselves as modern commercial robots, for example, iRobot’s Roomba (an autonomous vacuum cleaner).
There is no shame in utility: Hans Moravec, a pioneer in AI robotics from Carnegie Mellon's famed Robotics Institute, believes that finding practical domestic and industrial functions for robots will drive research, and aid in their becoming more intelligent , perhaps eventually bringing Asimov’s vision of a Humanity grandfathered-in by machine caretakers to fruition...
This site Copyright 2005 Mario Rodriguez. All rights reserved.