Thomas DeanAdjunct Professor of Computer Science
Contact InformationBox 1910
Providence, RI 02912
Email: tld at cs.brown.edu
Personal home page:
|CSCI1480||Building Intelligent Robots|
|CSCI1410||Applied Artifical Intelligence|
|CSCI0090-C||Talking with Computers|
Thomas Dean’s general research interests include automated temporal and spatial reasoning, planning, robotics, learning, and probabilistic inference. He is particularly interested in problems in which the notions of uncertainty and risk are complicated by the imposition of a limited time for deliberation and action. When the problems faced by an agent in a dynamic environment are complex, one typically must trade time for precision in making decisions: How long should you spend deciding what to pack when rushing to catch a plane? A similar tradeoff arises in deciding how much time and effort to expend in learning about a new or evolving environment: When visiting a new city, how much effort should you put into learning the streets and landmarks?
A more recent project involves building a probabilistic model of the neocortex. The basic tools for his research are derived from probability, statistics, Bayesian decision theory, and the design and analysis of algorithms. The applications involve mobile robots that perform such tasks as search and rescue, as well as and disembodied “knowbots” that operate on the World Wide Web.