Tuesday, May 03, 2005

Dilemmas of Prisoners and Professors

Some interesting game theory and philosophy from the last couple of NUMB3RS episodes. Usual spoiler warnings.

In the April 22nd episode Dirty Bomb there were three suspects who wouldn't talk. Charlie, the mathematician, likened the situation to Prisoner's Dilemma and suggested putting the suspects in the same room, which is usually the wrong thing to do in prisoner's dilemma. What Charlie did was compute the utility for each suspect cooperating (with each other and not the FBI) based on family considerations and their previous record and convinced the one with the most to lose by cooperating to defect and talk to the FBI. Clever, but I really wonder if that would work in real life.

Last Friday's episode Sacrifice took a more philosophical direction. A murdered think-tank computer scientist was developing a program that measured academic potential based on where someone grew up, down to a city block. If such a program actually worked, how should a program be used, if at all? How far should one go to stop the project?

Charlie and his physicist friend Larry ruminated on whether scientists are responsible for how their research gets used, as well as a discussion on the lonely life of a scientist at a lightly attended memorial service for the murder victim. The episode also had a physics joke I don't quite get.

Applied physicists are from Venus; Theoretical physicists wonder why it spins in the other direction.
I really enjoy those discussions between Charlie and Larry because they ask some interesting questions and add some dimension to a public view of mathematicians and scientists.

5 comments:

  1. Doesn't Venus rotate on its axis in the opposite direction of every other planet in the solar system? I think that is what the joke is getting at (or not, I'm no physicist either).

    ReplyDelete
  2. And the joke is also a play on the booktitle "Men are from Mars, Women are from Venus."

    ReplyDelete
  3. If such a program actually worked, how should a program be used, if at all? How far should one go to stop the project?

    Isn't that what SAT openly portends to be? a predictor of academic performance, based in as much data as possible, including cultural biases?

    ReplyDelete
  4. Venus' rotation is somewhat unusual in that it is both very slow (243 Earth days per Venus day, slightly longer than Venus' year) and retrograde.

    From here.

    ReplyDelete
  5. Yeah, Venus' slow rotation sucks from a terraforming point of view - even if you could get rid of the oppressive atmosphere, it would take a huge amount of energy to spin it up to a reasonable day (a large fraction of the total gravitational potential energy). Around a billion large comets falling in from the Oort cloud would be needed. Better just to use all of the Sun's radiant energy over the course of a year in order to disassemble it and build a Niven ring :-).

    ReplyDelete