Wen Dong (MIT, Media Laboratory),  “Modeling the Structure of Collective Intelligence” Advisors: Alex (Sandy) Pentland (2010) Human problem solving process can be described as a stochastic process of interacting agents. By observing many groups solving a wide variety of problems, we can determine the relationship between problem space structure and the solution process. To stochastically model human problem solution, this thesis presents the influence modeling, which attempts to capture the process by which an individual person navigates from one memory chunk randomly to another related memory chunk, and how a group of people randomly remind one another of memory chunks that could be individually. Using this method it is possible to account for about 50% of the performance variation between groups in terms of their discussion process without analysis of the underlying semantics. For details, click here.