I am a graduate student at Carnegie Mellon University pursuing my Ph. D. in Machine Learning. Currently, I'm working with Geoff Gordon on combining game-theoretic and first-order reasoning for multiagent games. I'm a member of the SELECT Lab.
Outside of the university, I'm interested in gaming (board and video), and science fiction (like any good nerd). I like learning about game design, particularly virtual world design and virtual world economics.
During my undergraduate career I've worked with Sandip Sen and the DAI Hards on multi-agent systems research; some results from my work can be found below.
I'm interested in multi-agent systems, logic, optimization, and game theory. Multi-agent planning, coordination, and inference are problems I'm especially interested in.
McDonald, Austin and Sandip Sen. "The Success and Failure of Tag-Mediated Evolution of Cooperation." Learning and Adaption in Multi-Agent Systems. Ed. Karl Tuyls. Berlin: Springer, 2006. 155-164. (pdf)
McDonald, Austin and Sandip Sen. 2004. Analyzing the Effects of Tags on Promoting Cooperation in Prisoner's Dilemma. In Artificial Multiagent Learning: Papers from the 2004 Fall Symposium, ed. Sean Luke, 9-15. Technical Report FS-04-02. American Association for Artificial Intelligence, Menlo Park, California. (pdf)
Available as pdf.
This is a compiler front-end for a Pascal-like language; it was completed as a class project in my Compiler Construction class. It's a fairly large project that's well documented and functional. It's written in C. (zip)
This is an NES emulator written in C++. It was a week-long project I attempted in order to gain a clearer understanding of console programming and hardware emulation issues. It's somewhat functional; sound code hasn't been included, and some of the harder bugs in the NES hardware haven't been fully replicated. It's available with test ROMs here (zip).
address:
5736 Hobart St Apt 3
Pittsburgh, PA 15217
phone:
918.344.7134
AIM: AustinM271
GTalk: mcdonald.austin