Current Projects

Geo-Spatial Enabled Dynamic Network Analysis, August 2007 - Present

The objective of this research is to develop geo-spatially enabled and constrained theory of social and knowledge networks.  A secondary objective is to understand and develop the technology, algorithms, and metrics needed to assess and visualize data that contain both geo-spatial and network information so that the proposed theory can be examined and validated. 

My work is principally aimed at a third objective: to understand how the inclusion of geo-spatial constraints will alter the predictions of dynamic network models, and whether this change increases the accuracy of the predictions. My thesis work dovetails nicely with this project as it sets a clear platform with which to test different methods of geo-spatially constraining models of network formation and dissolution.

Course of Action Analysis Using Integrated Micro and Macro Models of Adversarial Interaction and Activity (CAPES), June 2007 - Present

I am part of a team of graduate students and research staff in CASOS who are working on projects that explore the feasibility and value added of integrating different styles of micro level models – cognitive models, influence nets, Bayesian networks, actor based cost-benefit analysis with macro models of social and cultural change, resource based activity, and group processes.  This will involve linking social network models of information exchange and power with activity and decision making models at different levels of granularity (groups, political elite, population).  The integrated multi-level models will be used to explore issues of state failure and crisis mitigation and response.

Additionally, I am collaborating with a team from MIT, led by Dr. Mike Siegler, to integrate the RTE (a micro-level model) with one of their system dynamics model of state failure (a macro-level model). We hope to leverage the forecasts from one system for the other. For example, alliance networks, which the RTE is capable for making forecasts about can be used by MIT's model to help determine when fractionalized groups

Past Projects

Pre-Conflict Anticipation and Shaping (PCAS), February 2005 - September 2005

Under an SBIR with Aptima Inc., I led the development of Acumen, a dynamic-network multi-agent model of state stability. The goal of this project was to demonstrate the ability of computational models to assess nation-state stability and indicate possible policy options to mitigate risks of state failure.

In this role, I managed a software engineer, acted as the primary contact between the Carnegie Mellon team and the Aptima team, and represented the Carnegie Mellon team in meetings with DARPA every five weeks in Washington D.C. This work used VISTA as a point of departure and has led to the development of the RTE and my thesis work.

VISualization of Threats and Attacks (VISTA), January 2003 - January 2005

Working with Aptima Inc., we developed a proof of concept model of urban conflict for the Army which included a user interface. The goal was to enable an analyst to forecast conditions of interest in Military Operations other than War (MOOTW) and explore possible interventions to shape the environment. Aptima designed the user-interface and the team at Carnegie Mellon developed the underlying model. Results indicated that multi-agent model of urban conflict was possible, but it did point out a number of areas to focus on including constraints of data, issues of validation and evaluation, the importance of the user interface to help an analyst design hypothetical.

In developing VISTA, I acted as the team lead for Carnegie Mellon, managed two software engineers and helped present our progress at two meetings at the Army's Intelligence school at Ft. Huachuca, AZ.

Model-to-model comparison of OrgaHead and Simvision, January 2002 - June 2003

Before graduate school, I collaborated with the Stanford team to align OrgaHead and SimVision, two models of organizational design and evaluation. Docking serves three purposes, 1) to illuminate implicit assumptions, 2) to determine the points of similarity and differences in the models, and 3) to determine the ways that the models can be used in a complementary fashion. Results underscored how implementation differences between the models can affect their outcome measures.

Wireless Communications for Emergency Response in Allegheny County, August 2004 - December 2004