Thursday, December 16, 2010

Link mish-mash

All kinds of interesting stuff on the web today. Not that this wasn't also true yesterday, or the day before...

  • IBM says its Watson AI is ready to take on human Jeopardy champs for a $1 million prize. The showdown is scheduled for Feb 14. This is reminiscent of the Kasparov/Deep Blue showdown, except that Watson will be competing on human home turf: making sense of the linguistic ambiguity in the hints, phrasing, puns, etc. of Jeopardy prompts. (The AI has one advantage: I bet Watson will always remember to phrase its answer in the form of a question.)
  • A rehash of physicist Aaron Clauset's work on "the physics of terrorism." I'm not a big fan of his stuff, to be honest. My view: Clauset showed that the severity of terrorist attacks follow a powerlaw distribution, and has been wildly extrapolating from that single finding ever since.
  • A Jeremiad about the state of journalism, from Pulitzer prize winner David Cay Johnson. He talks trends (reporters know less and less about government; papers keep cutting content and raising prices), and hints at causes. The last paragraph is especially intriguing to me. Read it as a claim about how good reporting is supposed to uncover truth.
  • Predictions about 2011 from 1931. Eighty years ago, the NYTimes gathered a brain trust of experts in various fields and asked them what the world would look like today. Follow the link for predictions and some commentary. (Hat tip Marginal Revolution.)
  • This paper should depress my libertarian friends. Evidently, profit is evil, with an r-value of -.62.

2 comments:

Anonymous said...

Did you see this? Sure seems like Clauset has done a lot more than just one thing.

Abe said...

@Anonymous: Thanks for sending the link to Clauset's blog. I haven't read his latest papers, but a quick skim doesn't change my assessment too much. Instead of a one-liner and a link, let me give a more detailed description of where I stand.

As I read his work, Clauset has taken a data-centric approach to studying terrorism. As a social scientist, I applaud that direction. It makes good sense to use all the data available to look for patterns in human behavior, including terrorism. I'm certain that there are unknown, but discoverable laws governing the ways we interact. So far, Clauset and I agree.

My main criticism of Clauset is that he doesn't go far enough in untangling those laws. He's described some broad patterns (e.g. powerlaws in severity, increasing frequency over time), but doesn't have a specific, falsifiable theory of the causal mechanism behind those patterns. At best, he tells one story (one mechanism) that is consistent with his data. But he does not offer evidence to refute plausible alternatives.

Given that, it seems disingenuous to talk about how attacks are "highly predictable", and how "quantitative, model-based statistical forecasts" will allow us to combat terrorism. Clauset has some great new data and he's starting to untangle it, but he's overpromising when he suggests that he can predict violence in operationally useful ways.

In fairness, I don't know of anyone else who has built a theory of terrorism with this level of rigor. And it looks like Clauset is working in this direction, so my criticism may soften over time.

I especially like the suggestion that this work could eventually lead to an understanding of "fundamental constraints on terrorism that could serve as leverage points if exploited appropriately." That's real promise. It seems within reach, eventually. And when Clauset achieves it, I'll start promoting his work instead of sniping at it.

PS - If you are Aaron Clauset, it'd be nice to know it so I can stop referring to him/you by last name. Also, we could continue the conversation by email instead of debating in public.