Wen Dong focuses on modeling human interaction dynamics with stochastic process theory through combining the power of "big data" and the logic/reasoning power of state-of-the-art agent-based models, to solve our societies most challenging problems such as transportation sustainability and efficiency.
His approach is to drive "simulation games" (alike SimCity) with petabyte-sized "data exhaust trails" of networked people to study our social systems and make decisions through playing the data-driven simulation games on super-computers (which solve the stochastic differential equations that describe the agent behavior).
He has applied this approach to social systems of different sizes, ranging from the movement of millions of vehicle locations sampled every minute for years, the co-evolution of social network and individual behavior of different communities tracked by personal mobile phones and surveys, and Bales' interaction process analysis of several individuals solving problems tracked by socio-metric badges. The applications have won both industrial and academic awards.
Networking architectures, protocols, network control and management issues, and performance evaluation. Convergence of computer communications and telecommunications in WDM optical networks, mobile/wireless networks, the Internet, and other technologies like ATM and parallel and distributed processing. Peer-to-peer alternatives to traditional client-server models, ubiquitous computing with location-based services at the application-program level, mining for patterns in wireless and web, performance analysis of component models in application servers, and adaptive interfaces for collaborative systems. | More »
Computational theories for contour image analysis of things such as technical drawings, architectural plans, maps, and even cartoons to enable such images to be used in human and computer interaction. Design of computer chips to register a central region with high detail while displaying a larger zone at lower resolution. Algorithms that permit variable resolution image sequences to be understood. Techniques for visualizing common data-structures such as graphs and multidimensional data sets, and visualizing molecular structures. | More »
Methods and algorithms for putting data objects into categories. Methods of programming in which the machine is programmed to learn from an incomplete set of examples. | More »