The great race
The Fifth Generation: Artificial Intelligence and Japan's Computer Challenge to the World, by Edward A. Feigenbaum and Pamela McCorduck. Reading, Mass.: Addison-Wesley. 288 pp. $15.55.Skip to next paragraph
Subscribe Today to the Monitor
In the 4th century BC the Chinese philosopher Sun Tzu pointed out that knowledge is power. With it, he wrote, one can accomplish deeds surpassing all others.
His observation is particularly relevant today, as the race for dominance in Information Age technology speeds up, with the Japanese challenging the United States for the honor and the rewards of being first and best.
Ten years from now, if all goes according to schedule, 40 Japanese researchers and their flamboyant leader, Kazuhiro Fuchi, will introduce ''fifth generation'' computers, machines capable of intelligent reasoning that will be able, in effect, to troubleshoot problems of many kinds through factual and intuitive programs.
The impact of these machines on everyday work is expected to rival the changes wrought by the Industrial Revolution a century ago. Thus the country that introduces them stands to rise to world economic and strategic preeminence.
This is the prediction of authors Edward A. Feigenbaum and Pamela McCorduck. In ''The Fifth Generation: Artificial Intelligence and Japan's Computer Challenge to the World'' they urge the United States to respond more vigorously to the Japanese challenge.
Some American computer companies are also hastening to build fifth-generation computers, though without the kind of concerted, government-sponsored effort mounted by the Japanese.
Feigenbaum and McCorduck point out that the US has a two- to three-year lead in the necessary technology, but warn that ''we are squandering our lead at the rate of one day per day.'' They add, ''We are writing this book because we are worried.''
The concern is underlined in red throughout the book, which offers a fascinating and clearly written history of ''AI'' (artificial intelligence), along with the authors' views of its future. They also present an interesting sociological comparison of the Japanese and US approaches to research. The first several chapters praise Japan's accelerating progress in computer development and the careful research guidance provided by the country's Ministry for International Trade and Industry.
Although the Japanese are inexperienced in the crucial fields of software innovation and knowledge engineering, Feigenbaum and McCorduck reject the image of them as merely artful copiers. Feigenbaum, professor of computer science at Stanford University and one of the original developers of AI, notes that Kazuhiro Fuchi, who leads the Institute for New Generation Computer Technology (ICOT) in Tokyo, hired only young, promising researchers, contrary to Japanese tradition, and ''seems to have taken on a personal campaign to wipe out once and for all the energetic but uncreative stereotype that shadows the Japanese.''
The task the group at ICOT has undertaken is enormous. Whereas today's computers simply process data, tomorrow's will process knowledge. Rather than just store, manipulate, and retrieve facts, they will store relationships between facts in data bases compiled by experts in various fields. The machines will be capable of drawing inferences and reasonable conclusions. The effect will be similar to a car mechanic, say, deducing the problem with an automobile from a conversation with its owner. In essence, the new computers will serve as expert assistants to their users. They may usher in epoch-making change by providing immediate access to vast stores of knowledge and the ability to relieve users of many mundane tasks.
The question is: Can the Japanese produce the machines first?
While impressed with the Japanese determination and vision, Feigenbaum and McCorduck feel the US still has a strong position in the race. What they criticize is not American research ability but American dislike for group projects and wariness of any government role in a venture that is traditionally the preserve of private enterprise.
The authors conclude with a list of the options facing America: Do nothing, and muddle through; create industrial consortia to meet the challenge; form a joint venture with the Japanese; focus on software alone (which is America's strong point); or form a national lab for the promotion of knowledge technology. Feigenbaum and McCorduck strongly favor setting up such a laboratory.
Their book is highly useful for anyone interested in this crucial race.