Douglas Lenat Dead: AI Researcher Spent 40 Years in Building Computer With Common Sense

Douglas Lenat made a huge contribution to improving how computers operate. However, he wasn't able to finish the job.

Douglas Lenat Dead

Lenat, a pioneer in AI and a long-time researcher in computer science, has passed away. He passed away on Thursday in Austin, Texas. He was 72. According to his wife, Mary Shepherd, he died from bile duct cancer.

In an effort to replicate human judgment one logical rule at a time, Lenat spent over 40 years trying to teach computers common sense. Academics and programmers paid tribute to him online, praising his excellent career and determination in creating artificial general intelligence, or software that can reason.

Lenat graduated from the University of Pennsylvania with a bachelor's degree in mathematics and physics and a master's degree in applied mathematics in 1972. To earn his Ph.D., he moved to Stanford University, where he worked on software that can automatically create computer programs.

He later rose to the position of assistant professor at Stanford and Carnegie Mellon Universities, and he holds the unique distinction of having served on the scientific advisory boards of both Microsoft and Apple.

Lenat started Cycorp in 1994, a machine reasoning-focused AI startup, where he worked until his passing. Using a combination of a knowledge base and a reasoning engine, Lenat, a pioneer of neurosymbolic systems, attempted to train machines to reason. The system was used to power goods sold to businesses engaged in logistics and healthcare, and it had a natural language user interface.

ALSO READ: Robotaxis Unsafe or Safer Than Uber, Lyft Drivers? Here's What Experts Say

Lenat Developed AI Eurisko

Dr. Lenat created an A.I. while teaching computer science at Stanford University in the late 1970s. It was created to automate the discovery of new scientific concepts, methods, and laws through data analysis. He named the system Eurisko, a Greek term meaning "I discover."

Using a trillion-dollar budget, players in the 1981 role-playing game Traveller Trillion Credit Squadron designed and deployed a fleet of warships. He utilized this approach to evaluate the laws of the game. Go and Jeopardy are similar to chess. Later, the game was a great testing ground for cutting-edge A.I. technology.

After poring over the several volumes of the Traveller rulebook each evening, Eurisko discovered fresh strategies for winning the game. Some of them were absurd at one point; it implied that the only way to succeed was to alter the game's rules-but others showed promise.

Dr. Lenat would tweak the system every morning, steering it away from the absurd and toward the realistic. Eurisko eventually discovered an unconventional yet effective tactic under the guidance of his common sense.

Dr. Lenat participated in a Traveller tournament with several hundred other players over the Fourth of July weekend in nearby San Mateo, California. Eurisko's method helped him win the competition. The event's regulations were altered the following year to make the tactic ineffective. However, Dr. Lenat won the competition again after brainstorming with Eurisko to develop a fresh strategy.

The encounter gave rise to a fresh endeavor that would occupy him for forty years.

According to cognitive scientist Gary Marcus, he undertook an endeavor no one else dared to do. Although he failed, he had at least shown a portion of the route for those exploring the same path.