Artificial Intelligence

Artificial Intelligence

Shareece Barnes

CMGT/ 556

March 16, 2015

Andrew Nash

        Artificial Intelligence (AI), serves to mimic human thought processes and behavior using inanimate objects, such as computers, robots, etc. to perform resulting actions.  Intelligent systems are the applications of artificial intelligence and are comprised of the sensors, software and devices involved in human mimicry.  The ultimate goal of AI is to mimic human intelligence learning and developing as it encounters new experiences and data.  Although AI can be categorized in many ways, there are five major categories of artificial intelligence systems: expert systems, neural networks, genetic algorithms, intelligent agents, and virtual reality.  

The most common form of artificial intelligence is the expert system.  Expert systems are advisory programs that duplicate the complex reasoning skills of human experts to solve difficult problems.  They contain a knowledge base and rules for application of the knowledge base, varying upon the situation.   Applying expert systems expedite the business decision making process due to their ability to calculate various factors in a much shorter time frame than any human could ever manage.  In addition, information obtained by these systems is never forgotten as is the possibility with human beings.   Expert systems are commonly used to assist in medical diagnosis due to their ability to sort through a plethora of possible diagnoses based on the symptoms.

By contrast, a neural network analyzes a large amount of data to identify patterns and characteristics in situations where the rules are unknown.  It must adjust to the situation without a standard, therefore it must be capable of flexibility in its learning process.  Just as the expert system can be used to emulate an expert chess player by following the rules of the game to determine the best next move, a neural network uses the information provided to alert the business of oddities in the system.  This technology is used by many credit card companies to detect fraud, saving them an estimated $50 million per year.

A genetic algorithm uses a “survival of the fittest” approach.   Genetic algorithms can be used to eliminate options which may later prove to not be viable opportunities.    In this process, the system is optimized by constantly searching for better solutions to the problem at hand.  It generates the best output given the various inputs provided.  This system is useful for executives when planning what projects to pursue or discard for the most optimal advantage. It can also be used in determining technical improvements to the company, as is the case when telecommunications companies utilize genetic algorithms for configuring fiber-optic cables.  Needless to say, there is a wide variety of uses for genetic algorithms, as most things in business are at least comparable.