Artificial intelligence and fuzzy logic principles are also being applied to the development of neural networks use physical electronic devices or software to mimic the neurological structure of the human brain. The human brain has about 100 billion neurons or brain cells. However, these cells do not act as computer memory sites. No cell holds a picture of your dog or the idea of happiness. You could eliminate any cell or even a few million in your brain and not alter your mind. Where do memory and called synapses and they constitute about 40% of the brain mass. Using electrical pulses the neurons send on/off messages along the synapses. Computer-based neural networks work according to the same principle, using special AI software and complicated fuzzy logic processor chips to take inputs and convert them to outputs with a kind of logic similar to human logic.
Ordinary computers mechanically obey instructions according to set rules. However, neural network computers like children learn by example problem solving and memory by association. The network learns by fine-tuning its connections in response to each situation. ( in the brain learning takes place through changes in the synapses. ) Of your teaching a neural network to speak for example, you train it by giving it simple words and sentences, gradually change allowing more or less current to pass. Electronic neurons are primitive representations of real ones. However, neural networks are still startlingly intelligent one neural network learned how to pronounce a 20000-word vocabulary overnight. At first the network sounded like a child, groping for the right sounds then gradually mastering words and sentences.
Recently using software from a neural network producer Intel introduced a neural network chip that contains many more transistors than the Pentium, its newest microcomputer chip. Other chip makers are also working on neural network chips. Over the next few years these chips will begin to bring the Power of these silicon ‘’brains’’ not only your PC but also to such tasks as automatically balancing shifting laundry loads in washing machines. Many experts expect as information filtering tools on your PC and adjust to your needs. For example, they will sort your electronic mail in order of your priorities and automatically select online news items you would want and store them for your read later.
Here are two example of how neural networks have been used so far
> Using a PC and a $ 19 neural network program from California scientific software of Nevada city, California don Emmons a professional better recently picked the winning horse in 17 out of 22 races at a track in Detroit. He gathered data on the various horses past performances and fed it into the program.
> Montgomery county court house in Norristown Pennsylvania uses a neural network to save $40000 a years in fees for unneeded jurors with two previous year’s data on juror use and the next day’s trial schedule the system provides an immediate readout of the number of jurors required. The jurors who had been alerted 30 days ahead of time are called only if they are needed.