Research Interests

• Network Analysis -- Network representations permit a focus on structures of relationships rather than attributes of individuals. I'm particularly interested in networks formed by economic processes, such as supply chain relationships and logistical networks. The focus of my current research is on mathematical models that relate the structure of a network to the economic constraints under which it was formed.
  
  
• Probabilistic Graphical Models -- Many real world data sets describe places and times in addition to actors, attributes and relationships. PGMs are a useful framework for describing models that bridge all these types of data. My current research in this area is on unsupervised methods that reduce dependence on prior knowledge, such as the number of states that can be taken on by a hidden variable.
  
  
• Game Theory -- When considering interventions in economic systems, game theoretic models have the advantage of attempting to capture humans' ability to rationally adapt. Unfortunately, the complexity of real world systems and the limitations of human rationality make traditional game theoretic approaches inaccurate or intractable in many real-world situations. My research in this area has been towards relaxed solution concepts and compact game representations that hopefully take a step towards being able to apply these models to the types of systems analyzed in my other research.

Selected Publications

• Unsupervised Plan Detection with Factor Graphs. George B. Davis, Jamie Olson, Kathleen M. Carley. Sensor-KDD Workshop, ACM-SIGKDD, 2008.
  
  
• Factoring Games to Isolate Strategic Interactions. George B. Davis, Michael Benisch, Kathleen M. Carley, Norman M. Sadeh. Autonomous Agents and Multiagent Systems (AAMAS), 2007.
  
  
• Simultaenous Inference of Places, Activities and Behavioral Classes in Maritime GPS Data. George B. Davis, Kathleen M. Carley. Tech Report CMU-ISRI-07-113, 2007.
  
  
• Computational Analysis of Merchant Marine Data. George B. Davis, Kathleen M. Carley. Tech Report CMU-ISRI-07-109, 2007. Extended version.
  
  
• Algorithms for Rationalizability and CURB Sets. Michael Benisch, George B. Davis, Tuomas Sandholm.American Association of Artificial Intelligence Conference (AAAI), 2006. Extended version.
  
  
• Clearing the FOG: Understanding Interstitial Relationships with Fuzzy, Overlapping Groups. George B. Davis, Kathleen M. Carley.Social Networks, accepted for 2008 publication. Link is to working paper, not final version.
  
  
• Sequence Similarity Network Reveals Common Ancestry of Multidomain Proteins. Nan Song, Jacob M. Joseph, George B. Davis, Dannie Durand. PLOS Computational Biology 4 (5), 2008.
  
  
• Graph Theoretical Insights into Evolution of Multidomain Proteins. Teresa Przytycka, George B. Davis, Nan Song, Dannie Durand. Journal of Computational Biology 13 (2), 2006.
  

Education

• Intermediate Statistics, Fall 2006. Kathryn Roeder.
• Probabilistic Graphical Models, Fall 2006. Carlos Guestrin.
• Negotiation, Spring 2006. Linda Babcock.
• Foundations of Electronic Marketplaces, Fall 2005. Tuomas Sandholm.
• Graduate Artificial Intelligence, Spring 2005. Tuomas Sandholm and Mike Lewicki.
• Machine Learning, Spring 2005. Tom Mitchell and Carlos Guestrin.
• Fundamental Network Analysis Methods, (1/2 Semester) Spring 2005. David Krackhardt
• Electronic Voting, Fall 2004. Michael Shamos.
• Dynamic Network Analysis, Fall 2004. Kathleen Carley.
• COS Lab, Fall 2004. Latanya Sweeney.

Other Stuff

• I first became interested in networks as a computational biologist with Exelixis (2001-02) and the Durand Lab (2003).
  
  
• I enjoy constructive violence of all kinds, and studied and taught Taekwondo for a number of years. These days, I put more trust in good 'ol boxing and mixed-style grappling.
  
  
• I used to run the Informal COS Journal Club, and am currently looking for someone to take over some of the administrative load of resurrecting it.