Introduction and Background Fuzzy Logic Genetic Algorithms Neural Networks Feedback Links Page Information

Introduction and background

Anyone who has worked with computers eventually comes to the realization that there are some things the human mind can solve easily, but are virtually impossible for even the most powerful super-computer to compute. Artificial intelligence originally was developed during WWII as a code-breaking technique. Code-Breakers would create a large system of rules which would be followed by an expert system to simulate intelligence in code-breaking. In the past few years computer scientists, electrical engineers, and many others have realized that the same structure which makes the human brain so versatile can also be applied to solving problems in computing. In fact a great deal of arificial intelligence techniques derive their names from biological origins. For example genetic algorithims are based on the biological theory of evolution. Neural Networks came about from studying how neurons in the brain interact to process data and reach decisions. Human ability to deal with "fuzzy" data has also become a major topic of research with vast implications for the computing field. For example human ability to understand how much "a few" is rather than requiring a crisp number such as five.

One of the first people to devolp the theory of intelligent computers was the mathematician Alan Turing. Alan Turing was born on June 23, 1912 in london. As a boy Mr. Turing enrolled at the Sherbourne school in Deorset where he showed phenomenal ability in the field of mathematics. Around this time Mr. Turing became an atheist and began to wonder how the brain worked if there was no soul behind it. Mr. Turing believed that a machine could function in the same manner as a human brain and produce intelligent results. The machine Turing thought up was capable of reading a tape of infinite length. When the machine read the tape from left to right it would execute the command on the tape, much like computers of today read binary code and execute it. Mr. Turing also proposed that by altering the tape as part of the output the machine could "learn" from what it had done. From this Mr. Turing developed his famous "imitation test." This test was first published in Brain magazine in 1950. Mr. Turing's test was to place a person in contact with a computer and a human. If the person cannot determine which is the computer and which is the human by asking a series of questions, then the computer is thinking the same manner as a human.

Artificial Intelligence has not yet caught up with the sentient programs envisioned by science fiction writers and movie makers. However, Artificial Intelligence is used in today's world for everyting from medical applications to finances to computer games to speech recognition. Below is a list of applications that are in use or under development.

  • OCR or optical character recognition is one of the most extensively used application of neural network technology. This technology allows scanned documents to be converted from bitmaps into text files using neural networks.
  • Neural Networks have also been used in conjunction with fuzzy logic to enable hand-writing recognition.
  • Extensive work has been done using a variety of AI techniques to provide accurate financial forecasting. As of now, however, these techniques have proven only marginally effective.
  • Fuzzy logic has been used to more precisely control mechanical devices such as arie conditioners. Fuzzy logic is effective because it allows the device to do tasks such as cool the air "a lot" or "a little" or "none" rather than merely a binary choice of "cool" or "off."
  • Many of the speech recognition products on today's market use some form of AI to transcribe written speech.
  • Natural Language processing, which allows a user to ask the computer questions in ordinary english, utilizes some form of AI to process the request.

Sources

Gray, Paul. "Alan Turing." Time 29 Mar. 1999: 147.

Rao, Valluru B., and Rao, Hayagriva V. C++ Neural Networks and Fuzzy Logic. Henry and Holt Publishing Comp., 1995