| |
PAUL VARKEY
Multi Agent Systems Group |
|
Keywords interactive epistemology, games under incomplete information, decision-theoretic reasoning and optimal sequential planning under uncertainty, resource bounded rationality Overview My research interests can be best summarized in three words: uncertainty, epistemics and combinatorics. I am currently studying the world of agents, both, human and artificial. Eventually, I would also like to explore the natural world. In the world of interactions (both collaborative and strategic) between multiple (intentional and possibly rational) agents, (optimal) behavior ought to and oftentimes does depend on the knowledge of the agents. This knowledge may be incomplete, interactive (therefore, infinitely nested) and dynamic (continually evolving). A natural modeling tool for such epistemics is the probability calculus. The world of natural (ecological, geological, biological, atmospheric, etc.) phenomenon is seemingly rife with uncertainty or, at least, incompletely determined(able) causes, effects and correlations. Such phenomenon can sometimes be powerfully modeled and studied using probabilistic graphical models and the calculus of stochastic processes. In much of my studies, there is an underlying acknowledgement of the possibility that any mechanical solution procedure may have to evaluate an astonishing number of alternatives in the search for a solution. Therefore, I also believe it is important to consider how fast one can either solve a model completely or at least produce a sufficiently good enough solution. See conice resume (concise pdf, concise txt) and Research Statement (Mar 15, 2010, Dec 9, 2010). Papers
Talks
Posts
|
|