Artificial Intelligence

The term artificial intelligence (AI) refers to using computers to mimic intelligent processes found in the natural world. Researchers in AI have often overestimated their ability to produce intelligent behavior or thought processes with machines. However, AI has definitely had some success and will probably continue to exert greater impact on our society in the future. You only have to look at the latest robotic devices in an electronics store to glimpse some of the potential. We only briefly describe three of the major developments in AI: expert systems, neural networks, and genetic algorithms.

To build an expert system, you need a cooperating expert. An expert system tries to model the thinking process of someone who is an expert on a particular subject. For example, programmers might interview a doctor at length, trying to build various rules for decision making. For example, the doctor might say that the first thing to do is take the patient’s temperature. If the patient has a fever, you proceed along one set of steps. If the patient does not have a fever, you proceed along another set of steps. As you can imagine, the list of steps would be extremely long to diagnose many different illnesses. Some steps might lead to a dead end, where the system would indicate that a doctor would need to be consulted; the situation would be beyond the capability of the expert system. If you had an expert investor like Warren Buffett, you could potentially build a system that would mimic his stock selection process. However, researchers have found that it is extremely difficult to encode the decision-making process of many experts. Experts often seem to have a sixth sense in many situations that cannot be modeled.