June 13, 2005

Further to Heuristic Diversity

Posting will be light for the next few days, while I get various forms of work done. In the event you simply must have more of my prose to read, here is my latest (and probably last) article for the SFI Bulletin, which is the 2500 additional words on Scott Page's work on diversity I promised some time ago.

The tone is somewhat affected by the fact that the Bulletin is the functional equivalent of an alumni magazine for SFI --- it's supposed to make people feel good about the place and its work, and (hopefully) give money. Still, I think I did a pretty honest job of it. (On my own, I might have left out the bits about Scalia, and left in the bits about exactly what's wrong with The Wisdom of Crowds, but my editors were right on both counts.) Even at that length, I had to leave out a lot of things. The ones which I most regret, today, were the connections to boosting and other ensemble methods in machine learning and statistics, on the one hand, and to Charles Lindblom's work on decision-making by mutual adjustment and "disjointed incrementalism", a.k.a. "the intelligence of democracy". But I had the disturbing realization, when writing it, of just how easy it would be to turn it into a sixty- or hundred- page venture into social theory. Really, the subject demands a book; fortunately, Scott is writing one.

Now, back to the bowels of LaTeX.

The Collective Use and Evolution of Concepts; Self-Centered

Posted at June 13, 2005 12:13 | permanent link

Three-Toed Sloth