Artificial Intelligence (Ai)
The invention of the steam engine and industrial mass production astonished a society familiar only with manual labor, generating visions of a leisurely future in which human toil was no longer necessary and machines provided all of the necessities of life. Similarly, the power of digital computer technology to perform amazing feats of calculation has raised the hope of scientists that computers might be built which have the ability to perform complex tasks, make sound decisions in novel situations, initiate activity and learn from mistakes--in a word, to think . One of the most important figures in the early study of artificial intelligence was Alan Turing (1912-1954), an English mathematician and logician. Turing proposed a means by which one could determine whether a computer is "intelligent." Place a computer, Turing said, in one room and place a keyboard attached to the computer in another room. Then allow an individual to communicate with the computer via the keyboard. If the individual cannot determine whether there is a computer or a human in the other room, then the computer could be termed "intelligent." He argued that intelligent computers could eventually be created with the ability to play games like chess and checkers, learn and translate languages, create and break codes, and tackle tough mathematical problems. In the late 1930s Turing considered what mathematical procedures a computer would need to use in order to perform any calculation. Using a binary method, he illustrated on paper the operations by which a hypothetical computer could carry out any algorithm (a finite set of mathematical instructions). Turing's theorem was dubbed the " Universal Turing Machine" because it clearly demonstrated the properties that any genuine calculating machine would need to possess. Turing went on to help create a machine called Enigma which was capable of deciphering German codes during World War II. In 1956 an important conference on artificial intelligence was held at Dartmouth College. The Dartmouth symposium was the meeting place of a number of individuals who would shape the future of artificial intelligence including John McCarthy, a mathematics professor, Marvin Minsky (1916-), a Harvard mathematician interested in how the brain worked, Nathaniel Rochester (1919-), an IBM researcher, and Claude Shannon (1916-), a mathematician at Bell Laboratories. It was at this conference that Minsky began writing an influential paper eventually titled "Steps Toward Artificial Intelligence." In it he described several machines capable of proving mathematical theorems or playing chess. About the same time researchers at RAND Corporation completed work on an intelligent program. Allen Newell (1927-1992), Herbert Simon (1916-) and J. C. Shaw unveiled the Logic Theorist, capable of proving theorems in mathematics. It could carry out operations no one though a machine capable of performing, and it even provided some novel solutions to long-standing computer programming dilemmas. The Logic Theorist was the first device that really gave credence to the claim that computers could be made "intelligent." It was in the area of games that artificial intelligence gained international fame. IBM was responsible for two of the early game-playing programs. One of IBM's experts in vacuum tubes, Arthur Samuel (1901-1990), became fascinated with the idea of creating a checkers-playing program that would learn from its mistakes. He wrote a computer program that could recognize a position it had already come across in a previous game and alter its moves based on the results of the completed game.
In the late 1950s another IBM employee, Alex Bernstein, became one of the first to design a chess-playing computer program, but had limited success, as he encountered difficulty defining a clear set of principles that could guide the computer without testing every possible move. A team from Northwestern University eventually created a more successful chess program called CHESS. Their strategy was to "crunch" as many moves as possible; at any board position their program analyzes millions of possibilities. An early version of the program, CHESS 4.5, earned an expert rating from the United States Chess Federation. By 1978, their program could compete at the master level, though CHESS 4.7 lost to British master David Levy that year. Another chess-playing program called DEEP THOUGHT was developed by Feng-Hsuing Hsu at Carnegie-Mellon University. Capable of analyzing 34 million board positions per minute, DEEP THOUGHT beat several players at the master level before it was defeated by the world champion. In 1993 then next incarnation of DEEP THOUGHT was unveiled: DEEP BLUE. DEEP BLUE was capable of analyzing over 100 million board positions per second. In February of 1996, DEEP BLUE played chess champion Gary Kasparov and beat the computer four games to two. However, in a rematch the following year, DEEP BLUE beat Kasparov 3 1/2 games to 2 1/2. Although DEEP THOUGHT, and later DEEP BLUE, are highly successful at winning chess matches, a number of AI experts have become disillusioned with such programs, contending that they resemble a massive primitive calculators more than "thinking" machines. One of the most promising avenues of recent research in AI has been in language processing. One of the foremost innovators in the field is Raymond Kurzweil (1948-). After graduating from the Massachusetts Institute of Technology in 1974, Kurzweil began work on a reading machine for the blind that could translate any printed material into understandable speech. Kurzweil succeeded in producing such a device two years later. The Kurzweil Reader caught the attention of blind singer Stevie Wonder (1950-) who contacted Kurzweil, expressing interest in a keyboard synthesizer that could reproduce the sounds of acoustical instruments. In 1983 Kurzweil completed work on a groundbreaking synthesizer that could reproduce the sound of more than 100 musical instruments with surprising accuracy. He also began work on a voice-activated typewriter, a machine that IBM and other companies had been attempting to perfect since the 1940s. In the late 1980s Kurzweil produced Voice Works, a computer that can take dictation. He plans to apply the technology to create a writing machine that enables a deaf person to see a real-time readout of conversations and discussions. Other AI machines and programs have been designed to carry out foreign language translations, robot control, automatic programming, and natural language interaction. Recently developed expert system s can diagnose system malfunctions, develop manufacturing plant schedules, and even analyze chemical structures or disease symptoms.
This is the complete article, containing 1,043 words
(approx. 3 pages at 300 words per page).
