Sunday, 21 June 2009

Stuart Dreyfus on Mathematics, Chess, Expertise, Intuition, Intelligence and Artificial Intelligence


Stuart Dreyfus
Professor Emeritus, University of California, Berkeley
Ph.D. Applied Mathematics, Harvard University, 1964

Stuart Dreyfus is Professor Emeritus in Industrial Engineering and Operations Research at University of California, Berkeley. Stuart is also the younger brother of Hubert Dreyfus, philosophy professor at University of California, Berkeley. In 1986, the two brothers co-wrote Mind Over Machine: The Power of Human Intuition and Expertise in the Era of the Computer. In this book they discuss the 5 stages of skill acquisition (Novice / Advanced Beginner / Competent / Proficient / Expert). Stuart Dreyfus recalls "... what it is to think like a competent chess player, as he is stuck at that level."
I was always good at mathematics and took up chess as an outlet for that analytic talent. At college, where I captained the chess team, my players were mostly mathematicians and mostly, like me, at the competent level. At this point, a few of my teammates who were not mathematicians began to play fast chess at the rate of five or ten minutes a game, and also eagerly to play over the great games of the grandmasters. I resisted. Fast chess was no fun for me, because it didn't give me time to figure out what to do. I found grandmaster games inscrutable, and since the record of the game seldom if ever gave principles explaining the moves, I felt there was nothing I could learn from the games. Some of my teammates, who through fast chess and game studying acquired a great deal of concrete experience, have gone on to become masters.
It is interesting seeing this assessment by an accomplished mathematician, as it does reinforce the point that being good at mathematics does not necessarily imply increased chess expertise. That being said, mathematicians do seem to be over-represented in the youthful, and also not so youthful, chess playing fraternity. On Stuart Dreyfus' Berkeley webpage their is also the following quote:
"Expertise is pattern discrimination and association based on experience. It is intuitive. There is no evidence you can reduce it to rules and theory. Hence, Artificial Intelligence probably can't be produced using rules and principles. That's not what intelligence is."
I now become confused when considering the chess skill of a computer program such as Fritz or Rybka. Clearly these programs are expert at playing chess but it is not the same as the expertise of a human chess player. The expert human chess player is very intuitive in their play, whereas the expert computer chess program is very heavily analytical in it's approach. Maybe we need a different set of terms for the skilled expert computer program.

References:
Mind over machine: The power of human intuition and expertise in the era of the computer.
Hubert L Dreyfus (Author), Stuart E. Dreyfus (Author)
Hardcover: 231 pages
Publisher: Free Press; First Edition edition (1986)
ISBN-10: 0029080606
ISBN-13: 978-0029080603
[Amazon link]