CONTACT: Stanford University News Service (650) 723-2558
STANFORD -- In 1984, a Japanese modular-house manufacturing company, Sekisui Heim, learned about an offshoot of artificial intelligence (AI) research called expert systems. The company bought a majority interest in a small AI company and invested about $3.5 million in applying the new technology to improve the efficiency of their manufacturing process. As a result, the company estimates it is now saving about $8 million annually.
According to Edward A. Feigenbaum, professor of computer science at Stanford University, the Sekisui story illustrates the way in which the Japanese have mastered expert- system technology and integrated it into their industrial operations as a core competence.
"There don't seem to be a lot of barriers to the introduction of expert-system technology in Japan. Companies just do it without endless analysis and discussion," Feigenbaum said.
He recounted the story as part of a verbal summary of the report "Knowledge-Based Systems in Japan" that he gave to the annual meeting of the American Association for the Advancement of Science in San Francisco on Wednesday, Feb. 23. Feigenbaum chaired the panel assembled by the Japanese Technology Evaluation Center at Loyola College in Maryland that prepared the analysis. The report was funded by the National Science Foundation and the Defense Advanced Research Project Agency and released last May with little fanfare.
Expert systems are computer programs that capture human expertise to assist people in performing a variety of tasks, including diagnosis, planning, scheduling and design. These systems have become the most successful commercial applications of artificial intelligence (AI) research. Thousands of expert systems are now in routine use worldwide by business, industry and government.
In the area of expert-system research, the panel found that the U.S. retains the lead in both quality and quantity of basic research at universities and industrial laboratories. In applied research and development, the U.S. and Japan support roughly the same quantity of work but the Japanese have taken the lead in quality.
In terms of applications of the new technology, however, the quality and quantity (relative to each country's gross domestic product) of Japanese systems have reached parity with American systems, and the number is growing more rapidly than in the United States. In addition, the Japanese now lead in terms of the integration of expert systems with other software, its use in consumer products, and the support system for development of new systems, the panel concluded.
"As far as applications go, the Japanese are good and getting better, while the situation in the United States is good but getting worse," Feigenbaum said.
One reason for the difference is that such large Japanese computer companies as Hitachi and Fujitsu have embraced expert- system technology and can provide software and technical support for other companies that want to utilize it. In America, however, the large computer makers offer little software and almost no support for expert systems. Most of this support is coming from smaller software companies, he said.
Although expert-system technology has spread through a wide segment of industry and government in both the United States and Japan, there are some notable contrasts. In particular, the Japanese emphasize expert-system effort in the steel and construction industries, while in the United States the emphasis is on financial services, manufacturing and aerospace.
This is an archived release.
This release is not available in any other form.
Images mentioned in this release are not available online.
Stanford News Service has an extensive library of images, some of which may be available to you online. Direct your request by EMail to firstname.lastname@example.org.