Apr 24, 2019 - Technology

Remembering an AI pioneer

Photo of a man sitting in front of a blackboard

Nilsson in 1987. Photo: Ed Souza / Stanford News Service

Nils Nilsson, a pioneer in artificial intelligence whose groundbreaking work helped lead to today's online maps, died yesterday at the age of 86.

The big picture: Nilsson's work was foundational to several modern technologies. Anyone who has gotten directions online, played a video game or marveled at moving robots has experienced the legacy of his research.

  • Nilsson developed a famous algorithm called A* for finding the shortest path from one point to another, while working at the Stanford Research Institute, now SRI International.
  • He helped build Shakey, the first robot that could plan a complex route on its own. Starting in 1966, Shakey wobbled through the halls of SRI by relying on a map and various sensors — much like today's autonomous cars do.
  • He wrote a textbook in 1980 that helped shape a generation's understanding of his field.

Nilsson's navigation work "motivated everything that's in Waze and Google Maps," says Frank Chen, a partner at Andreessen Horowitz who studied under Nilsson at Stanford. "That's pretty lasting influence."

In his 2010 history of AI, Nilsson wrote: "I have participated in the quest for artificial intelligence for fifty years — all of my professional life and nearly all of the life of the field."

  • His quest included the exciting dawn of AI, as well as a period known as the "AI winter" during which investment and interest in the field dried up. That's when Chen took Intro to AI with Nilsson. He was one of about 100 students, he tells Axios — a tiny number compared with the several thousand who today take introductory AI classes at Stanford every quarter.
  • In an oral history recorded in 1989 — winter's coldest hour — Nilsson was undeterred:
"A lot of people in computer science would say, 'Well, AI's intractable, because the problems are all exponential. … ' My point of view about that is that that might be. Posed in that fashion, they may be intractable. [But] people do them."

Now, decades later, one of his biggest contributions is experiencing a resurgence. After a period of unpopularity, a method Nilsson championed — giving an AI system specific rules, rather than having it learn them from data — is thought to be able to take AI to the next level, if paired with new techniques like deep learning.

The breadth of Nilsson's legacy has been reflected on Twitter:

  • Andrew Ng, a Stanford AI researcher, wrote: "RIP to my friend, colleague, and AI visionary Nils Nilsson. Your work on the A* algorithm has improved countless lives. … I will always remember your work, but even more importantly your kindness."
  • Rodney Brooks, the robotics pioneer, wrote: "My 1972 introduction to AI research was his 1971 book "Problem-Solving Methods in AI". Thank you Nils!!! RIP."
  • Judea Pearl, the leading AI researcher at UCLA, called Nilsson "an AI pioneer, and a mentor to many of us since the 1970s. Always encouraging and always insisting on understanding new ideas, and how they fit together in the grand scheme. I will miss him immensely."
Go deeper