Recently, the media has spent an increasing amount of broadcast time on new technology. The focus of high-tech media has been aimed at the flurry of advances concerning artificial intelligence (AI). What is artificial intelligence and what is the media talking about? Are these technologies beneficial to our society or mere novelties among business and marketing professionals? Medical facilities, police departments and manufacturing plants have all been changed by AI but how? These questions and many others are the concern of the general public brought about by the lack of education concerning rapidly advancing computer technology. Artificial intelligence is defined as the ability of a machine to think for itself. Scientists and theorists continue to debate if computers will actually be able to think for themselves at one point (Patterson 7). The generally accepted theory is that computers do and will think more in the future.
AI has grown rapidly in the last ten years chiefly because of the advances in computer architecture. The term artificial intelligence was actually coined in 1956 by a group of scientists having their first meeting on the topic (Patterson 6). Early attempts at AI were neural networks modeled after the ones in the human brain. Success was minimal at best because of the lack of computer technology needed to calculate such large equations. AI is achieved using a number of different methods.
The more popular implementations comprise neural networks, chaos engineering, fuzzy logic, knowledge based systems, and expert systems. Using any one of the aforementioned design structures requires a specialized computer system. For example, Anderson Consulting applies a knowledge based system to commercial loan officers using multimedia (Hedburg 121). Their system requires a fast IBM desktop computer. Other systems may require even more horsepower using exotic computers or workstations. Even more exotic is the software that is used. Since there are very few applications that are pre-written using AI, each company has to write it’s own software for the solution to the problem. An easier way around this obstacle is to design an add-on. The company FuziWare makes several applications that act as an addition to a larger application. FuziCalc, FuziQuote, FuziCell, FuziChoice, and FuziCost are all products that are used as management decision support systems for other off-the-shelf applications (Barron 111). In order to tell that AI is present we must be able to measure the intelligence being used.
For a relative scale of reference, large supercomputers can only create a brain the size of a fly (Butler and Caudill 5). It is surprising what a computer can do with that intelligence once it has been put to work. Almost any scientific, business, or financial profession can benefit greatly from AI.
The ability of the computer to analyze variables provides a great advantage to these fields. There are many ways that AI can be used to solve a problem. Virtually all of these methods require special hardware and software to use them. Unfortunately, that makes AI systems expensive. Consulting firms, companies that design computing solutions for their clients, have offset that cost with the quality of the system. Many new AI systems now give a special edge that is needed to beat the competition. Neural networks have entered the spotlight with surprisingly successful results.
A neural network is a type of information processing system whose architecture is similar to the structure of biological neural systems (Butler and Caudill 5). The neural network tries to mimic the way a brain and nervous system work by analyzing sensory inputs and calculating an outcome. A neural network is usually composed of simple decision making elements that are connected with variable weights and strengths. Each one these elements is called a neurode. The term neurode is similar to the biological neuron. The term was modified slightly to indicate an artificial nature. Memory is stored by a certain pattern of the connection weights between the neurodes. Processing information is performed by changing and spreading the connection’s weights among the network. Before it can be used a neural network it must be trained.
Some can learn by themselves, some require training by doing, and others learn by trial and error. A computer learns by naturally associating items the computer is taught and grouping them together physically. Additionally, a computer can retrieve stored information from incomplete or partially incorrect clues. Neural networks are able to generalize categories based on specifics of the contents. Lastly, it is highly fault tolerant. This means that the network can sustain a large amount of damage and still function. Its performance fades proportionally as the neurodes disappear (Butler and Caudill 8). This type of system is inherently an excellent design for any application that requires little human intervention and that must learn on the go. Created by Lotfi Zadeh almost thirty years ago, fuzzy logic is a mathematical system that deals with imprecise descriptions, such as “new”, “nice”, or “large” (Schmuller 14). This concept was also inspired from biological roots. The inherent vagueness in everyday life motivates fuzzy logic systems (Schmuller 8).
In contrast to the usual yes and no answers, this type of system can distinguish the shades in-between. In Los Angeles a fuzzy logic system is used to analyze input from several cameras located at different intersections (Barron 114). This system provides a “smart light” that can decide whether a traffic light should be changed more often or remain green longer. In order for these “smart lights” to work the system assigns a value to an input and analyzes all the inputs at once.
Those inputs that have the highest value get the highest amount of attention. For example, here is how a fuzzy logic system might evaluate water temperature. If the water is cold, it assigns a value of zero. If it is hot the system will assign the value of one. But if the next sample is lukewarm it has the capability to decide upon a value of 0.6 (Schmuller 14). The varying degrees of warmness or coldness are shown through the values assigned to it. Fuzzy logic’s structure allows it to easily rate any input and decide upon the importance. Moreover, fuzzy logic lends itself to multiple operations at once. Fuzzy logic’s ability to do multiple operations allows it to be integrated into neural networks.
Two very powerful intelligent structures make for an extremely useful product. This integration takes the pros of fuzzy logic and neural networks and eliminates the cons of both systems (Liebowitz 113). This new system is a now a neural network with the ability to learn using fuzzy logic instead of hard concrete facts. Allowing a more fuzzy input to be used in the neural network instead of being passed up will greatly decrease the learning time of such a network.
Another promising arena of AI is chaos engineering.
The chaos theory is the cutting-edge mathematical discipline aimed at making sense of the ineffable and finding order among seemingly random events (Weiss 138). Chaologists are experimenting with Wall Street where they are hardly receiving a warm welcome. Nevertheless, chaos engineering has already proven itself and will be present for the foreseeable future. The theory came to life in 1963 at the Massachusetts Institute of Technology. Edward Lorenz, who was frustrated with weather predictions, noted that they were inaccurate because of the tiny variations in the data. Over time he noticed that these variations were magnified as time continued. His work went unnoticed until 1975 when James Yorke detailed the findings to American Mathematical Monthly. Yorke’s work was the foundation of the modern chaos theory (Weiss 139). The theory is put into practice by using mathematics to model complex natural phenomena.
The chaos theory is used to construct portfolios of long and short positions in the stock market on Wall Street. This is used to assess market risk accurately, not to predict the future (Weiss 139). Unfortunately, the hard part is putting the theory into practice. It has yet to impress the people that really count: financial officers, corporate treasurers, etc. It is quite understandable though, who is willing to sink money into a system that they cannot understand? Until a track record is set for chaos most will be unwilling to try, but to get the track record someone has to try it, it’s what is known as the “catch-22.” The chaos theory can be useful in other places as well. Kazuyuki Aihara, an engineering professor at Tokyo’s Denki University, claims that chaos engineering can be applied to analyzing heart patients. The pattern of beating hearth changes slightly and each person pattern is different (Ono 41). Considering this discovery a dataprocessing company in J! apan has marketed a physical checkup system that uses chaos engineering. This system measures health and psychological condition by monitoring changes in circulation at the fingertip (Ono 41). Aihara admits that chaos-engineering has tremendous potential but does have limitations. He states, “It can predict the future more accurately than any other system but that doesn’t mean it can predict the future all the time.” Along these lines Rabi Satter, a computer consultant with a BS in Computer Science, believes that the current sentiment that the world is rational and can be reduced to mathematical equations is wrong. “In order to make great strides in this arena [AI] we need new approaches informed by the past but not guided by it. A fresh voice if you would. As one person said we are using brute force to solve the problem” states Satter.
A few more implementations of artificial intelligence include knowledge-based systems, expert systems, and case-based reasoning. All of these are relatively similar because they all use a fixed set of rules. Knowledge-based systems (KBS) are systems that depend on a large base of knowledge to perform difficult tasks (Patterson 13). KBS get their information from expert knowledge that has been programmed into facts, rules, heuristics and procedures. However, the power of a knowledge-based system is only as good as the knowledge given to it. Therefore, the knowledge section is usually separate from the control system and can be updated independently. This enables system updates and additional information to be added in a more efficient manner then making a whole new system from scratch (O’Shea 162). Expert systems have proven effective in a number of problem domains that usually require human intelligence (Patterson 326). They were developed in the research labs of universities in the 1960’s and 1970’s. Expert systems are primarily used as specialized problem solvers. The areas that this can cover are almost endless. This can include law, chemistry, biology, engineering, manufacturing, aerospace, military operations, finance, banking, meteorology, geology, and more. Expert systems use knowledge instead of data to control the solution process. “In knowledge lies the power” is a theme repeated when building such systems.
These systems are capable of explaining the answer to the problem and why any requested knowledge was necessary. Expert systems use symbolic representations for knowledge and perform computations through manipulations of the different symbols (Patterson 329). But perhaps the greatest advantage to expert systems is their ability to realize their limits and capabilities. Case-based reasoning (CBR) is similar to expert system because theoretically they could use they same set of data. CBR has been proposed as a more psychologically plausible model of the reasoning used by an expert while expert systems use more fashionable rule-based reasoning systems (Riesbeck 9). This type of system uses a different computational element that decides the outcome of a given input. Instead of rules in an expert system, CBR uses cases to evaluate each input uniquely. Each case would be matched to what a human expert would do in a specific case. Additionally this system knows no right answers, just those that were used in former cases to match. A case library is set up and each decision is stored. The input question is characterized to appropriate features that are recognizable and is matched to a similar past problem and its solution is then applied.
Now that each type of implementation of AI has been discussed, how do we use all this technology? Foremost, neural networks are used mainly for internal corporate applications in various types of problems. For example, Troy Nolen was hired by a major defense contractor to design programs for guiding flight and battle patterns of the YF-22 fighter. His software runs on five on-board computers and makes split-second decisions based on data from ground stations, radar, and other sources. Additionally it predicts what the enemy planes would do, guiding the jet’s actions consequently (Schwartz 136). Now he and many others design financial software based on their experience with neural networks. Nolen works for Merrill Lynch & Co. to develop software that will predict the prices of many stocks and bonds. Murry Ruggiero also designs software, but his forecasts the future of the Standard & Poors index. Ruggiero’s program, called BrainCel, is capable of giving an annual return of 292%.
Another major application of neural networks is detecting credit card fraud. Mellon Bank, First Bank, and Colonial National Bank all use neural networks that can determine the difference between fraud and regular transactions (Bylinsky 98). Mellon Bank states the new neural network allows them to eliminate 90% of the false alarms that occur under traditional detection systems (Bylinsky 99). Secondly, fuzzy logic has many applications that hit close to home. Home appliances win most of the ground with AI enhanced washing machines, vacuum cleaners, and air-conditioners. Hitachi and Matsu*censored*a manufacture washing machines that automatically adjust for load size and how dirty the articles are (Shine 57). This machine washes until clean, not just for ten minutes. Matsu*censored*a also manufactures vacuum cleaners that adjust the suction power according to the volume of dust and the nature of the floor. Lastly, Mitsubishi uses fuzzy logic to slow air-conditioners gradually to the desired temperature. The power consumption is reduced by 20% using this system (Schmuller 27). The chaos theory is limited in scope at this time mainly because of lack of interest and resources to experiment with. However, Wall Street will be hearing more about it for a long time to come. Also, the medical field has an interest because of its ability to distinguish between natural and non-natural patterns. The chaos theory has a foot in the door, but a breakthrough in design will have to come around first before any major moves toward the chaos theory will happen. Expert systems are prevalent all over the world. This proven technology has made its way into almost everywhere that human experts live. Expert systems even can show an employee how to be an expert in a particular occupation.
A Massachusetts company specializes in teaching good judgment to new employees or trainees. Called Wisdom Simulators, this company sells software that simulates nasty job situations in the business world. The ability to learn before the need arises attracts many customers to this type of software (Nadis 8). Expert systems have also been applied in medical facilities, diagnosis of mechanical devices, planning scientific experiments, military operations, and teaching students specialized tasks. Knowledge-based systems and case-based reasoning will be on the rise for a long time to come. These systems are souped-up expert systems that provide more powerful searching and decision-making strategies. KBS is finding its home at help desks by working with telephone operators to direct calls. CBR will have close ties to law with its ability to use past precedents to determine a sentence and prison term. KBS is already being used by the Tennessee Department of Corrections for determining which inmates are eligible for parole (Peterson 37). Making recommendations on which AI systems work the best almost requires AI itself. However, I believe that some are definitely better than others. Neural networks, unfortunately, have performance spectrums that continue to dwell at both extremes. While there are some very good networks that perform their designed task beautifully, there are others that perform miserably. Furthermore, these networks require massive amounts of computing resources that restrict their use to those who can afford it. On the other hand, fuzzy logic is practically a win-win situation. Although some are rather simple, these systems perform their duties quickly and accurately without expensive equipment.
They can easily replace many mundane tasks that others computer systems would have trouble with. Fuzzy logic has enabled computers to calculate such terms as “large” or “several” that would not be possible without it (Schmuller 14). On the other hand, the chaos theory has potential for handling an infinite amount of variables. This gives it the ability to be a huge success in the financial world. It’s high learning curve and its primitive nature, however, limits it to testing purposes only for the time being. It will be a rocky road for chaos theory and chaos engineering for several years. Finally, expert systems, knowledge-based systems, and cased-based reasoning systems are here to stay for a long time. They provide an efficient, easy to use program that yields results that no one can argue with. Designed correctly, they are can be easily updated and modernized. While the massive surge into the information age has ushered some old practices out of style, the better ones have taken over with great success. The rate of advancement may seem fast to the average person, but the technology is being put to good use and is not out of control. A little time to experiment with the forefront technologies and society will be rewarded with big pay-offs. Soon there will be no place uncharted and no stone unturned. Computers are the future in the world and we should learn to welcome their benefits and improve their shortcomings to enrich the lives of the world.
Bibliography:
Bibliographies/ Works Cited
Barron, Janet J. “Putting Fuzzy Logic into Focus.” Byte April 1993: 111-118.
Butler, Charles, and Butler and Caudill, Maureen. Naturally Intelligent Systems. Cambridge: The MIT Press, 1990.
Bylinsky, Gene. “Computers That Learn By Doing.” Fortune 6 Sep. 1993: 96-102.
Liebowitz, Jay. “Roll Your Own Hybrids.” Byte July 1993: 113-115.
Hedburg, Sara. “See, Hear, Learn.” Byte July 1993: 119-128.
Nadis, Steve. “Artificial Intelligence.” Omni August 1993: 8.
Ono, Takashi. “Chaos is Coming.” World Press Review April 1993: 41.
O’Shea, Tim. Intelligent Knowledge-Based Systems. Cambridge: Harper & Row, 1987.
Patterson, Dan W. Introduction to Artificial Intelligence and Expert Systems. Englewood Cliffs: Prentice Hall, 1990.
Peterson, Ivars. “From Prisons to Autos to Space.” Science New 17 July 1993: 37.
Riesbeck, Christopher K., and Schank, Roger C. Inside Case-Based Reasoning. Hillsdale: Lawrence Erlbaum Associates, 1989.
Schmuller, Joseph, ed. PC AI March/April 1993: 8.
Schmuller, Joseph. “Three Faces of Fuzziness.” PC AI March/April 1993: 14-15, 26-29.
Schwartz, Evan I. “Where Neural Networks are Already at Work.” Business Week 2 Nov. 1992: 136-137.
Shine, Jerry. “Fuzzy Visions.” Popular Science April 1993: 57.
Weiss, Gary. “Chaos Hits Wall Street-The Theory, That is.” Business Week 2 Nov. 1992: 138-140.