Artificial Intelligence
By : DR. ALIA A. AL-JINDI
SCHOOL OF EDUCATION KING FAISAL UNEVERSITY K.S.A.
ABSTRACT
The intention of this paper is not to describe the sophisticated intricacies of the technical aspects of artificial Intelligence. Instead, the primary objective is to introduce the concept of Artificial Intelligence and its current status along with future expectations. Social ramifications and global competition will also be included.
One hundred years ago, the presence of a nonhuman entity approaching or exceeding man's intelligence would be impossible. Today, however, technology has brought us far enough to say that such intelligence does exist in certain forms. Artifical Inteligence (AL) is a branch of computer science that imitates human thinking, problem-solving, reasoning, and understanding of natural language(3). Examples of Al that are in existence today involve computers that can recognize your name. They can read articles and write summaries of them. Some of these machines exist only in laboratories, but others are functioning in practical situations(4).
In spite of this definition, people have been arguing for many years whether or not computers are intelligent machines. One of the reasons is that there is confusion over a definition of intelligence itself. Human intelligence is so diverse that it is extremely difficult to define and measure(5). This statement makes some people believe it is impossible for computers to exhibit human intelligence(9). Others claim, however, that it is possible to build a computer in a way that will come close to the working capacity of the human brain. Their claim depends on the test used to measure the mental ability of humans including tests for abstract reasoning, verbal reasoning, numerical reasoning, and space relationship and perception.
These test results are the then compared to the computer's results from the same test areas. The computer rates higher in some of these areas than in others, but is improving in almost every way. However, regardless of the words used in trying to define Al, such a phenomenon is definitely a reality today and the foundation of tomorrow's supercomputers, who know how far this vehicle for information will take us ?!
The debate now focuses on how we can determine if a machine can think. Turning (1950) suggested that we could measure a machine's thinking by its ability to fool us under cross-questioning into mistaking it for a human being one example of such a test was performed on a program called DOCTOR designed by Joe Weizenbaum. The test involved a computer conducting a psychiatric interview where patients were questioned through the medium of teleprinter. In the end, 60% of the participants though they had been communicating with a real live doctor(8).
The history of Al began in the early 19th century with Charles Babbage, who is considered to be the father of computers. He developed a machine called an "analytical engine". This machine, which is generally agreed upon to be the ancestor of our modern digital computer, was never built. At the time people questioned whether this "dream" machine could actually think. One of Babbage's colleagues, Countess Lovelace, wrote an essay describing the analytical engine and said "no" it could not think(2). From that "unexisting" machine, the concept of computers developed. If you look at the different generations of computers, the first generation is charcterized by being powered by vacuum tubes. The development of transistors in 1948 marked the second generation of computers. In 1959 the third generation came about because of the integrated curcuit chip invention.
This led to the fourth generation of supercomputer which has even more power and ability. It is predicated that by 2001 the compactness and reasoning power of computers will match that of the human brain(5). All of these steps make it possible for us to have the "Expert System". Expert systems are computer programs that can copy expert knowledge within a specific field to solve selected problems(7). Computer specialists and experts of a given field design expert systems by expoloring the reasoning used to apply knowledge in that certain subject. Expert systems reach conclusions through human processes like reasoning(5). The expert system typically consists of three different systems combined to function as one. The input/output system feeds data into the computer and retrieves advice while the second system, the inference system, uses reasoning of human experts. The third system, the knowledge acquisition system, enters information into the system. The information exits the system via the input/output system.
Appendix a illustrates the flow of information through the expert system(2). DENDRAL was one of the first programs using an expert system that was thoroughly accepted by most people to be working example of an application of Al in a real life situation. DENDRAL helps organic chemists determine the molecular structure of unkown compunds. Based on the fact that there are many expert systems at work in various professions throughout the world already, predicted applications for these expert systems unlimited(6). After DENDRAL, a group of Carnegie - Mellon University researchers built a knowledge - based system called HERSAL and HARPY to understand continuous human speech. The vocabulary was limited. The system moved along slowly, but in their limited ways they worked and provided some useful ideas about improving the future(2). The largest single group of expert systems is centered around medicine. Harry Paple, a computer scientist, and Jack Meyers, a physician, created INTERNIST/ CADUCEUS at the university of Pittsburgh. This expert system does diagnoses in internal medicine at an expertise level. It serves as a great diagnostic ad. The INT / CAD covers more than 80% of all internal medicine, has a knowledge base of about 500 diseases, and more than 3,500 manifestations of disease(2). With the MYCIN system at Standford University, the physician supplies the patient history and laboratory test results to the system and lets the program reason out possible diagnoses. The system diagnoses blood and meningitis infection and then advises the physician on antibiotic therapies for treating the infection. If the physician is uncertain about the diagnosis, he will question the program as its reasoning(2). Another expert system, the Ventilation Management Assistant, was developed at the Pacific Medical Center. It provides advice to clinicians about patients undergoing mechanical ventilation. A ventilation is a piece of medical equipment that assists in helping a critically ill patient to breathe (2). There are many other expert systems in existence for medical aid. Systems are in use for digitalis dosing, glaucoma diagnosis and treatment, renal disease, arthritis and rheumatism, fetal illnesses, and even the development of new drugs(2). Beside the field of medicine, the expert system is also used in the field of human communication activities such as summarizing research, writing speeches, monitoring plitical campaigns, and evaluating instructional strategies(11). The future of Al is seen in light of many different aspects. Speed is a great concern, for the more speed a computer has, the faster Al will expand. The
Japanese are expected to improve today's machine speed 10 to 100,000 logical inferences per second (LIPS) to 100 million to a billion LIP(2). Computer power is also a critical issue when considering Al. On the fundamental level, the first tool that is needed to support computer development is computer power. Parallelism, the belief that processors can run in parallel as compared with today's method of single tunnel processing, allows much more power(1), Allen Newell of Carnegie-Mellon says "You go to parallelism because you want more power"(7). Language development is yet another concept researchers are working on understanding as part of their goal to make computers "smarter". In a few years a keyboard might not be necessary for computers will respond to spoken commands(10). Progress that has been made in the last few years encourages the idea that a new era in both talking and listening computers is dawning. Computers are learning to handle language more and more on a basis. This will have serious implications on the evolution of Al, for language is one of the signs of intelligence as we know it. In 1981 the fifth Generation Project was initiated by the Japanese. They claim that by 1992 they will develop computers that will infer about and respond coherently to information given them visually or orally(7). Chip design has important bearing on the future of Al. Chips are getting smaller and smarter. Gallum Arsenide is replacing silicon as chip making material because it is so much faster. Electrons travel through Gallum Arsenide almost 5 times faster than they do through silicon(10). A subset of the chip is the Biochip. Eventually, the Biochip will be developed to replace the Gallum Arsenide chip. This Biochip will be built along biological lines rather than from non-organic substances. It will be built from complex organic chemicals and be known technically as an "organic data processor". Compared to the silicon chip's holding 50,000 to 100,000 bit of information, a Biochip holds 50 billion to 100 billion bits. This is a remarkable millionfold increase. To show how tiny biochips are, one trillion biochips could fit on a postage stamp. The power of biochip is awesome, operating 10 billion times faster than the most advanced home computer in use today(10). The social implications involved with such computer transformations are dramatic. The computer will no longer be sophisticated data manipulators but instead intelligent assistants. As an example, in education the implementation of Al systems will redefine the teacher's role. Students will have access to virtually any source of information in the world through satellite connected networks(3).
Of course, along with the good, comes the bad which would include human dependence on an increasingly computerized urban system. There is govenment backing in some countries for computer research. Japan and the United States lead the race for computer research even though Japan has strong governmental committments in terms of Co-operation, resources, and funding. The Japanese government is currently funding a project to develop a parallel interface machine along with a knowledge based machine and then merge the two to form a Fifth Generation Computer(4). In conclusion, Artificial Intelligence design is struggling through its developmental stages like any other emerging technology. In time it will have thinking abilities equivalent to our own. Computers are immune to distractions that complicate man's reasoning finding logical solutions. With wisdom and a careful attitude, which is not of total dependence, we will reap the full benefits from artificial intelligence.
REFERENCES
1. Edelson, E. (1984). Computers at the forge. Mosaic, 1, 261-277.
2. Feigenbaum, E., & McCorduck, P. (1983). The Fifth Generation.
Massachusetts : Wesley.
3. Green, J. O. (1984). Artificial intelligence and the future classroom.
Classroom Computer Learning, 4 (6), 26-31.
4. Highberger, D. (1984). Intelligent computing era takes off. Computer
Design, 23 (10), 79-95.
5. Hyde, M. O. (1982). Computer that think. New Jersey : Enslow.
6. McCorduck, P. (1979). Machines who think. San Francisco : W. H.
Freeman.
7. Metzger, No. (1984). Two find the way forward. Mosaic, 15, (10,2-9.
8. Nichie, D. (1982). Machine intelligence and related topics. New York :
Gorden and Breach.
9. Neil, G. (1979). Artificial intelligence. Blue Ridge summit, PA: Tab Books.
10. Ritchie, D. (1984). The binary brain : Artificial intelligence in the age of
electronics. Boston: Little, Brown.
11. Skopec, E. (1984, March). Expert systems and human communication.
Paper presented at the Annual Meeting of the Eastern Communication Association, Philadelphia.
إرسال تعليق على: "Artificial Intelligence"