Artificial Intelligence, Expert Systems, Neural Networks

Group 10

Cameron Kinard, Leaundre Zeno, Megan Wiedmaier, Heath Carley

Key Words: Artificial Intelligence, expert systems, neural networks, neurons

Technology has a huge impact on the world we live in today. It helps organize and manage the large amount of information used in daily activities. Information Technology (IT) is any computer-based tool that people use to work with information and support the processing needs of an organization. Information technology (IT) is becoming increasingly important for organizations to incorporate into their decision making process. A good IT system is vital to the operations of an organization. Artificial Intelligence (AI) has long been a powerful tool commonly used in many IT systems. AI has been developed into many different forms. Each form has its advantages and its disadvantages, but when properly designed, it can be a valuable asset to any organization. AI expands on our human brain power to support decision making. Artificial Intelligence (AI) is being used in many different industries including the military, banks, hospitals, retail, credit card companies, and airline companies.

Artificial Intelligence (AI) is a branch of science dealing with behavior, learning, and adaptation in machines. AI is generally divided into two general categories: conventional artificial intelligence and computational intelligence. Conventional artificial intelligence is a method involving the use of structured formulas and statistical analysis. Conventional artificial intelligent methods include expert systems, case based reasoning, Bayesian networks, and behavior based AI. Computational intelligence is the method of analyzing existing information and recognizing patterns. Simply put, it has the ability to learn from existing information. Computational intelligence methods include neural networks, evolutionary computation, and fuzzy systems. Two of the most common AI methods are expert systems and neural networks. These two systems have been developed through AI research and are sometimes discussed as separate entities from AI. We will discuss expert systems and neural networks as components of AI.

Expert systems were first developed in artificial intelligence during the 1960s and the 1970s and were absorbed commercially in the 1980s. The most common form of an expert system is a program structured by a set of rules and procedures that take the knowledge, supplied by an expert, and recommend a course of action in order to solve specific problems. Expert systems use reasoning to work through problems and offer recommendations that address these problems. They are ideal for diagnostic and prescriptive problems. Expert systems are usually built for specific applications called domains. They are used for accounting, financial management, production, process control, medication prescriptions, and in many other domains. One of the strengths of an expert system is the gathering and use of expertise. This strength is a competitive advantage for organizations that have a defined set of policies and procedures. Expert systems can perform many functions benefiting organizations such as reduction in training costs, decrease human error, providing consistent answers to repetitive tasks, and safeguard sensitive company information. On the other hand, they also have a few disadvantages. One drawback is the expert’s inability to organize his knowledge into a useable formula that can consistently and accurately solve problems. The expert system also has the disadvantage of not being able to solve problems for which it was not designed. A significant drawback of an expert system is its inability to use common sense and judgment to solve newly encountered problems.

Neural networks are artificial intelligence systems that can be trained to recognize patterns and adapt to new concepts and knowledge. Unlike expert systems, neural networks are not bound by a set of rules designed for a specific application. These networks are able to imitate the human ability to process information without following a set of rules. A neural network uses interconnecting neurons to produce an output. The output relies on the collaboration of individual neurons, and the processing of information is done in parallel. Basically, a neural network uses its neurons collectively to execute its functions. A neuron is the basic functioning element in a neural network that takes inputs and produces outputs. This allows the neural network to continue performing even if some of its neurons are not functioning. Neural networks are most useful for identification, classification, and forecasting when dealing with a large amount of information. They are used in speech and visual recognition. The proper use of these systems can be very beneficial to a business. They can give an organization a competitive edge by decreasing cost and increasing efficiency. These systems are widely used in engineering, drilling, meteorology, medical, insurance industries and by the military. Neural networks have many advantages. They can adjust to new information on their own. Neural networks are able to function without structured information. They are able to process large volumes of data. The biggest disadvantage of neural networks is the hidden layers. The fact that these layers are hidden prohibits users from adjusting the connections reducing control of the system.

Many organizations use these systems to improve their operations. The systems increase completion rates and decrease error by reducing human interaction. These systems protect information and utilize knowledge more efficiently to make intelligent decisions. Companies can gain an edge over their competitors by implementing these systems. Many different industries use these systems in a variety of applications. Banks use the systems to organize operations, to project stock activity, and to manage properties. Hospitals use them to develop work schedules, to track patient room assignments, and to diagnose and treat illnesses. They are used in the detection of fraud by credit card companies. One of the more interesting ways they are being used is in robotics. Manufacturers have been using robots for many years, but with the advancement in neural network technology, they have become increasingly important in industrial operations. Robots often perform jobs that are too dangerous for humans. They are especially effective doing repetitive jobs. The medical field is now developing micro robots that can be injected into a patient’s arteries or veins to perform a complicated surgery.

These intelligent systems are the future of industry. They have and will continue to revolutionize the way the world functions. It has already been proven that a properly designed system have already been used to enhance an organization’s operations. The uses of these systems seem to have no limits, and with creative people behind the technology, we may one day see a true creative machine.

References

Wikipedia. (2006). Artificial intelligence. Retrieved November 26, 2006, from http://en.wikipedia.org/wiki/Artificial_intelligence

Wikipedia. (2006). Expert systems. Retrieved November 26, 2006, from http://en.wikipedia.org/wiki/Expert_systems

Wikipedia. (2006). Neural network. Retrieved November 26, 2006, from http://en.wikipedia.org/wiki/Neural_network