My research, using the ten hundred most common words

A while ago, I ran across the famous XKCD comic, Up Goer Five.  That comic inspired the namesake text-editor, which makes for a very interesting communicative challenge.  Being a fan of unique ways to communicate science to the masses (e.g. The Tiny Transactions of Computer Science), I took the time to write down what my thesis work aims to achieve using the ten-hundred most common words:

Computers can tell stories, but they can’t see what goes on inside your head; more to the point, a computer can’t see what you remember, which is actually a key part of how well you enjoy a story. My work deals with finding out what you remember from a story as you are reading it. I will then get a computer to track what you remember, so it can change the story on the fly to give you a great one. My work can be used in both usual stories, and in games.

I submit this to Ten Hundred Words of Science, and I encourage you to do the same.

Game-Theoretic Analysis of Doing What You Like v. Getting Tenure

Earlier this month I had a chance to sit for lunch with a longtime friend and professional colleague, Dr. David L. Roberts.  We had a small chat of the State of the Union, where I commented on one of the personal weaknesses I have when it comes to professional engagements: I have trouble saying ‘no’. (I’m working on it!)

In general, I like doing a lot of things and I get distracted by shiny objects.  David smiled the smile of “been there, done that,” and said: “let me show you something that was shown to me in the context of tenure.”  Using a really bad paper napkin and a pen that could barely get ink out, he drew the following chart:

(Doing…)	Things that get you tenure	Things that don’t get you tenure
Things that you like	[ +2, +2 ]	[ +1, -1 ]
Things that you don’t like	[ +1, +2 ]	[ -1, -1 ]

Game-Theoretic Analysis of Tenure – Your payoffs are shown in bold, with the other payoffs corresponding to the University, which we will assume wants you to get tenure and has the capacity to do so.  +2 payoff is highly desirable, +1 payoff is desirable, and -1 payoff is undesirable.  The strategy of doing “Things that you like” strictly dominates doing things you don’t like.

The justification for the payoffs is simple: you prefer doing things you like all the time, you’d rather get tenure than not, and each individual reward is independent and discrete (so getting tenure is +1 unit of reward and doing something you like is another distinct +1 unit of reward).

Anyway, the point he made (and very eloquently), is that “doing things you like is a strategy that strictly dominates doing things you don’t like”.  Recalling the wonders of dominated strategies:

We say that s_i strictly dominates s′_i if, for every choice of strategies by the other players, the payoff to player i from using s_i is greater than the payoff to player i from using s′i.

I felt it was a very interesting game theory problem, since you can kind of abstract this setup to anything that fits the pattern: “things you want to do v. things you have to do”.  In fact, David brought it up precisely because I had a situation that fit the pattern: the problem of doing things you like v. graduating with a Ph.D.  The above abstraction does not capture (even though it could) the repercussions of taking extra time in doing things you like that don’t get you tenure or get you graduated, so if you would like to do this for your own case, be sure to change the payoffs to your desired level of detail and/or sophistication.

I think I’ll be chasing some more shiny objects until these payoffs change for me.