2009 Poster Sessions : Society Design and Cooperative Agents

Student Name : Michael Munie
Advisor : Yoav Shoham
Research Areas: Artificial Intelligence
How to aggregate the knowledge and beliefs of a group of agents is a difficult but fundamentally important problem. We will look at three techniques to make these decisions. First, imagine we have a small group of agents trying to decide if they should work independently and collaborate in the end, or work collaboratively the whole time -- which is the most effective way of working? The results here show that this problem is much more difficult that we might imagine. Second, we look at the well-known problem of voting. We quantify the nature and frequency of these problems and find voting to be much better in practice than suspected. Finally we look at the situation where there is a large society of agents with diverse abilities (like in a corporation) working to solve a specific problem (of a large and interconnected nature) and ask ourselves: how should we organize society?

Michael is a Ph.D. student at Stanford working with Professor Shoham in the multiagent systems group. He graduated with degrees in Mathematics and Computer Science from the University of Illinois at Urbana-Champaign. He has worked at Lockheed Martin on prediction markets and information aggregation and at Fermilab on pseudorandom number generation. His current work focuses on how societies can work together to make optimal or near optimal decisions.