Big Presentation
Yesterday, I gave a presentation to Stuart Russell, Daphne Koller, and Andrew Ng, some of the biggest names in Artificial Intelligence. Stuart Russell is one of the coauthors of the cannonical AI text book, Artificial Intelligence: A Modern Approach. Daphne Koller is in the process of writing a book on Probabilistic Models and is one of the foremost researchers in Bayesian and Markov processes. Andrew Ng is a Machine Learning researcher, probably best known for making a helicopter fly upside down. Of course, they couldn't have cared less about my presentation; I was demoing a system called Gamemaster, which is the server for General Game Playing and this year's evaluation tests for the Transfer Learning project, which they will most likely not be using. Either way though, it was cool to be up in front of the room giving a talk to some of the most distinguished professors and text book authors in the field of AI.
A little background for those who are interested: General Game Playing is the area of study of creating computer programs that can play any game given to them, where a game is formally defined by its rules. The idea is that the computer should be able to figure out the best strategy from the rules alone. Transfer Learning is the idea that if an agent (or person) plays a game, and then plays a different but similar game, they should do better on the second game because after playing the first because they learned something useful by playing the first game. My research project involves creating the formal mathematics which define how that transfer occurs as well as setting up systems for other research groups to test their machine learning agents on.
blog comments powered by Disqus