Scientists to Learn About ‘Smart’ Computers

Computer scientists and managers from some of the nation’s largest companies are gathering in Claremont this summer to learn how to create computer systems that can, in the words of the Pentagon, “see, hear, talk, plan and reason.”

Under the auspices of the Los Angeles-based Institute of Artificial Intelligence, about 250 professionals are paying $495 to $5,875 for the opportunity to study with authorities in the field from Carnegie-Mellon University, Stanford, Duke and the Rand Corp., among others.

Scientists from corporations including Lockheed, Hughes Aircraft and Litton Industries are among those devoting themselves to the study of machine vision, logic programming and knowledge engineering. The sessions at Harvey Mudd College offer a comprehensive look at artificial intelligence, or AI, which depends on a new generation of super-powerful computers programmed to solve complex problems in a way that closely mimics the human thought process. The most common use is in so-called expert systems, which collect information from human experts and are then programmed to solve problems.

“We wanted to figure out a way to train managers and technicians to put together an expert system,” said Jonathan Simonds, president and co-founder of the 6-month-old institute. “We don’t presume to train someone with zero knowledge of computer science.”


Expert systems have been designed to provide practical assistance to a range of professions from doctors to oil drillers. One system helps doctors diagnose blood-related diseases and prescribe treatment. Schlumberger’s “Dipmeter Adviser” analyzes oil well drilling statistics and helps find underground oil reservoirs.

Simonds said the institute seeks to create a “critical mass to keep artificial intelligence alive.” Its goal for the inaugural summer session is to produce about 150 newly skilled computer engineers who can go back to their companies and put their knowledge to immediate use.

There is already some skepticism about the sophisticated systems because the first commercial AI industry fell flat on its face in the 1960s after a string of false starts and overly ambitious promises. But Simonds said he fears that if companies do not invest in AI, the field may go through a “winter.”

The privately funded institute has received about $500,000 from individuals and companies and Simonds said there is another $500,000 “sitting on the sideline” for future use.


“We are trying to make this discipline profitable for the people and companies who want to invest in it,” he said.

In the past two years, a majority of America’s 100 largest corporations have been studying how artificial intelligence can be applied to their business, according to a recent report by Montgomery Securities in San Francisco.