Machine Learning

Parallelism computing architecture would allow the computer to do things in real time in architecture not possible to do at all. GPS, CAs, neural nets, and expert systems would penetrate all areas and no area would be immune from eventual exploration; information projects will be done on multiple fronts at once; and most of the work would not be done by computer scientists or Knowledge based people but most likely by those in end-use fields, who see a need and have a more-or-less direct contact with these more theoretical sources of ideas. GPS, neural nets, CAs, Gas, and expert systems are a commodity for everyone to use. Natural language semantic networks will become more available. Simon seems optimistic that semantic networks accuracy will improve, as a he says, “I think our main bottlenecks there today are not in natural language, but in our ability to mount projects big enough to grow the data bases that would be necessary to do a job on the full language with all of its possibilities for metaphor.” ---