Adam Campbell

From Santa Fe Institute Events Wiki

I am currently a Computer Science Ph.D. student at the University of Central Florida. My main research interests are cooperative multi-agent systems and evolutionary computation, and I'd like to use a complex systems' approach to studying both of them. I also enjoy programming competitions such as TopCoder and ACM's ICPC, but have had little time to do either in the past couple years. I look forward to meeting everyone and seeing the diversity of research in Santa Fe.

Answers to Daniel Rockmore's questions:

  1. What are your main interests?
    • My main research interests are in distributed Artificial Intelligence; specifically, cooperative multi-agent systems. Along with trying to get groups of agents to work together to solve tasks, I am also interested in modeling and understanding both natural and man-made multi-agent systems with abstract computer simulations. Aside from computer science research, I enjoy reading about general science, especially theoretical physics and evolutionary theory.
  2. What sorts of expertise can you bring to the group?
    • Over the past decade or so, I have written many multi-agent programs such as cellular automata and systems containing many agents coordinating their actions. I usually use the MASON simulation framework. I also have a pretty good knowledge of genetic algorithms and neuroevolution (evolving artificial neural networks).
  3. What do you hope to get out of the CSSS?
    • To understand multi-agent systems better, I need to understand the theory behind complex systems, emergence, and self-organization. I look to gain a better mathematical understanding of these concepts from the CSSS. Besides these technical goals, I look forward to meeting fellow graduate researchers to see the current research that is being done with regards to complex systems.
  4. Do you have any possible projects in mind for the CSSS?
    • Currently, it is difficult to evolve behaviors for a team of agents because it is difficult to quantify teamwork. How do we know that what an agent did at time X led to the team completing the goal at time Y? It would be interesting to see if there is a way to analyze the evolved behaviors of an agent in order to determine what the role of that agent is in a team. I wonder if the work done in Cosma Rohilla Shalizi's PhD thesis on causal architectures and statistical complexity can be used to address these issues.