Artificial Intelligence Just Lost a Leader
Plus a position-search announcement from NSF
Roger Schank just passed away. Roger was a top leader of AI. I overlapped with him for my time at Yale. In 1974, he became a professor of computer science and psychology at Yale University. In 1981, Schank became Chairman of Computer Science at Yale and director of the Yale Artificial Intelligence Project. I was gone by then off to Berkeley.
John Brockman wrote: He is a computer scientist and cognitive psychologist who has worked in the AI field for twenty-five years. Like Marvin Minsky, he takes the strong AI view, but rather than trying to build an intelligent machine he wants to deconstruct the human mind. He wants to know, in particular, how natural language— one’s mother tongue — is processed, how memory works, and how learning occurs. Schank thinks of the human mind as a learning device, and he thinks that it is being taught in the wrong way. He is something of a gadfly; he deplores the curriculum-based, drill-oriented methods in today’s schools, and his most recent contributions have been in the area of education, looking at ways to use computers to enhance the learning process.
Roger’s View
Roger was a scary guy—especially for junior theory faculty like me. I did get along with Roger and he did support me when I was up for tenure. But he was scary no matter what. Part of the issue was that Roger did not think much of any research that was not AI based. Theory was not a top issue in his view.
Beat Roger Out
At Yale once Schank walked into Alan Perlis’s office and told Alan that he had thought about some post-doc position we had funding for. Schank began to explain to Perlis that he, Roger, had a couple of ideas of who he planned to hire to fill this position—of course the position was his. Perlis waved him off and said,
Lipton already has found a candidate and the position is gone.
I was not there, so all this is second-hand from Perlis, but I can imagine that Roger was not pleased—in his mind all resources of the department should have gone to build up the AI group. Roger was initially upset, but after this event he acted differently toward me. He treated me with more respect—in general theory was not his type of research. I always thought that he respected someone who “beat” him at the resource game, since he was such a strong player. I probably should not do it again, but doing it once was cool.
Years later, after Roger had moved to Northwestern University, he tried hard to hire me. Perhaps I should have gone. Oh well.
Open Problems
We want to announce this from Dilma Da Silva:
The Division Director, Computing and Communication Foundations (CCF) of NSF: NSF/CCF is looking for an IPA (rotator) Program Director (PD) for the Algorithm Foundation cluster. The job posting for the IPA PD position is available at here. My colleague Tracy Kimbrel and I will be happy to address any questions that potential applicants may have.
[sourced Brockman quote, added subtitle for NSF announcement]
Roger and I were “friends” — kind of “New York friends” — since we met at the Stanford AI project in late 69 or early 70 — so about 53 years or so. He was really great at thinking things out from scratch — like an ancient Greek — and some of his insights really helped make progress.
I won’t mention the names, but when I buttonholed one of Minsky’s students at an AI conference and complained that his kewl whizz-bang logic hacking mechanism he was presenting didn’t have anything to do with the way people did things he said “That’s what Minsky said, it doesn’t matter.” But Minsky signed his thesis. Roger wouldn’t sign theses he didn’t believe were actually contributing to AI.
Roger put on a scary guy act, but he was actually a nice guy. He was a big football fan, and when he’d show up late to a game (of touch) his students were playing, he’d figure out which team was losing, and coach/advise them. Also, when Roger left Yale, he busted his butt to make sure as many of his remaining students as possible finished. For the brief period I was at Yale (all-but-thesis, ’84*), I never saw him do anything other than do everything he possibly could for his students.
*: After passing the quals, I wasn’t finding a thesis topic I wanted to work on, didn’t really have the energy/interest at that point in my life, and punted. That’s on me, not Roger. Also, I saw the writing on the wall that doing AI would mean spending my life screaming at idiots, and I still think I was right about that. (Roger was, as am I, unimpressed with the current round of AI. Long story short: John McCarthy got it right in 1956: AI should be the study of human cognition using computation (the mathematical abstraction) to inform that study and programming actual computers to verify the results of that study. That’s why, when recently asked about the current state of AI, Roger said: “There’s no such thing as AI.” )
Perhaps, there must be such a thing as AI.
My distant memories are of a lecturer’s seeming avoidance of the nomenclature. He kept calling that thing Expert System. It is not easy for one to come up with such an intelligent mnemonic for the notion, named after some artifacts.
The following looks quite artificial (ASCII not representing octets). Maybe, it carries some intelligence?
Intelligence can not be artificial and anything artificial just can not be intelligence. We seem to have to choose one.
Anybody can offer us an easy pick?
Sorry. There is always surprises with different handling of HTML or LaTeX at different places, which is hard to know before hand.
Sorry. Didn’t know in blockquoted blocks, wrapping around can stop working.
— RH