Work, AI
Work has been somewhat less stressful the past couple of weeks. I put in some time this week to finish an old project. We'll see Monday whether I did it right or not.
My little AI foray has taken many (MANY!) twists and turns, but things are progressing nicely. I figured out a way to get my algorithm(s) to handle datasets in the millions of elements, rather than the thousands. This means it's far more practical. (In CS-speak, I figured out how to reduce the space complexity of the algorithm from quadratic to linear without sacrificing its linear time complexity.)
It turns out Netflix is holding one of those $1M contests to make an algorithm that predicts how people will review movies.
EDIT (03-11-2008):
Contrary to what I thought I'd read, Netflix does, in fact, include temporal data with their testing database. So, it's possible the AI thing I've been dinking with could find a use for this.
My little AI foray has taken many (MANY!) twists and turns, but things are progressing nicely. I figured out a way to get my algorithm(s) to handle datasets in the millions of elements, rather than the thousands. This means it's far more practical. (In CS-speak, I figured out how to reduce the space complexity of the algorithm from quadratic to linear without sacrificing its linear time complexity.)
It turns out Netflix is holding one of those $1M contests to make an algorithm that predicts how people will review movies.
EDIT (03-11-2008):
Contrary to what I thought I'd read, Netflix does, in fact, include temporal data with their testing database. So, it's possible the AI thing I've been dinking with could find a use for this.