Douglas Lenat, an artificial intelligence researcher who spent nearly 40 years trying to build common sense into computers, recreating human judgment one logical rule at a time, died Thursday in Austin, Texas. He is 72.
His wife, Mary Shepherd, said the cause was bile duct cancer.
In the late 1970s, as a professor of computer science at Stanford University, Dr. Lenat developed an AI system he called Eurisko — a Greek word meaning “I discovered.” It is designed to automate the discovery of new scientific concepts, methods and laws through data analysis.
In 1981, he used this system to study the rules of a very complex role-playing game called Traveler Trillion Credit Squadron, where players use a trillion-dollar budget to design and deploy a fleet of warships. like chess, You go and Jeopardy! In the coming years, the game will be a perfect place for the latest AI technology.
Each night, after combing through multiple volumes of the Traveler rulebook, Eurisko identified new ways to win the game. Some are ridiculous — at one point, it suggested the best way to win is to change the rules — but others are promising.
Every morning, Dr. Lenat the system, pushing it away from the ridiculous, towards the practical. Guided by his common sense, Eurisko eventually finds an unusual but powerful strategy. Instead of spending the trillion-dollar budget on large, mobile, protected warships — as other players have done — it proposed building hundreds of small ships that are largely immobile and poorly protected but carry enormous firepower.
At the end of July 4, Dr. Lenat at a Traveler tournament in nearby San Mateo, Calif., competing with several hundred other players. Using Eurisko’s strategy, he won the tournament. The next year, the tournament organizers changed the rules so that the strategy no longer worked. But after working with Eurisko to discover a new strategy, Dr. won again. Lenat in the contest.
The experience inspired a new project that would consume him for the next four decades.
By running on dozens of computers, Eurisko could discover possibilities that Dr. Lenat — and other people. But it needs help from human judgment. Machines cannot be truly intelligent, he realized, unless they also have common sense.
The project is called Cyc. He sets out to identify the basic but largely unspoken laws that frame how the world works, including everything from “you can’t be in two places at once” to “when drinking a cup of coffee, you hold open ended.” He knew it could take decades – maybe centuries – to complete the project, but he was determined to try.
In recent years, the Cyc project – and the rule-based approach to AI research it represents – has fallen out of favor with leading AI researchers. Instead of defining intelligence rule by rule, line of code by line of code, the giants of the technology industry are now focusing on systems that learn skills by analyzing massive amounts of digital data. This is how they developed popular chatbots like ChatGPT.
Many leading researchers now believe that this type of extensive data analysis will eventually breed common sense and reasoning. But while today’s computers struggle with even simple tasks and play fast and loose with reality, others believe the industry could learn from Dr. Lenat and his endless struggle to build common sense by hand.
“These chatbots think that when you drive a nail into the wall, it should be vertical,” said the Northwestern University professor and AI researcher. Ken Forbes. “They can be very useful. But they don’t understand the world.”
Douglas Bruce Lenat was born on September 13, 1950, in Philadelphia, a son of Nathan and Gertrude (Cohen) Lenat. When he was 5 years old, he and his family moved to Wilmington, Del., where his father, a trained chemist, owned a bottling company called London Dry.
After his father’s death in 1963, he returned to the Greater Philadelphia area with his mother, and his older brother Ronald. As a high schooler in Wyncote, Pa., his after-school job was cleaning goose cages and rat cages. To find a better life, he learned to program computers.
At the University of Pennsylvania, he completed three degrees in four years — a bachelor’s degree in math and physics and a master’s degree in applied mathematics — before moving to the West Coast for his doctorate. He enrolled at Stanford to study artificial intelligence. His thesis committee included three of the researchers who founded the field in the late 1950s.
It was an unpredictable period for artificial intelligence research — later dubbed AI Winter. But Dr. Lenat is among a new generation of researchers who have revived interest in what has been a decades-long struggle to create machines that can mimic the brain.
In the early 1980s, some of the nation’s top tech companies helped create a corporation meant to keep the United States at the forefront of technological research: the Microelectronics and Computer Technology Corporation, or MCC headed by Admiral Bobby Ray Inman — former director of naval intelligence, former director of the NSA and former deputy director of the CIA — the corporation hired Dr. Lenat as its chief scientist in 1984. From the company’s new headquarters in Austin, Texas, he began work on his common sense engine.
Two years later, he told Time magazine that the project would require 350 man-years of work to reach success.
In 1994, when another AI winter arrived, Dr. Lenat the project to a new company called Cycorp. Funded by various government organizations and private companies, he continued to develop his common sense engine until his death. He and his collaborators spent more than 2,000 man-years on the project, writing more than 25 million rules.
In addition to his wife and brother, Dr. Lenat has a daughter, Nicole Danielle Hermanson, from his first marriage, which ended in divorce; and two grandchildren.
An inveterate traveler, Dr. Lenat has visited over 100 countries and all seven continents. After the cremation, Ms. Shepherd, he plans to fix some of his debris scattered across the moon.
In the fall, as ChatGPT captured the public imagination, Dr. Lenat and cognitive scientist Gary Marcus in a new paper meant to show a new generation of researchers what they can learn from his nearly 40 years of work at Cyc. While working on the project, he had a recurrence of cancer that first appeared in 2021.
In July, Dr. Dr. Lenat Marcus to help him finish the paper. An abridged version was published a month before his death. “He took on the project that no one else had the courage to do,” said Dr. Marcus. “He did not completely succeed. But he showed us at least part of the way.
