CS 4702: Artificial Intelligence: Uncertainty and Multi-Agent Systems

Cross-Listing: None

A key issue in the design of intelligent system is how to deal with uncertain or incomplete information, as obtained, for example, through (noisy) sensory input. The first half of this course focuses on how to represent and reason with uncertain information. The second half covers the study and design of multi-agent systems. Topics include Bayesian networks, dynamic Bayesian networks, belief propagation, Markov random fields, exact and approximate probabilistic inference methods, Markov Chain Monte Carlo methods, connections to statistical physics and information science, adversarial reasoning and planning in multi-agent systems, and game theoretic notions underlying multi-agent systems. This course complements COM S 472, but is given as a self-contained unit.

Offered: Spring only

Prerequisites: CS/ENGRD 2110 and CS 2800 or equivalent.

Grade options: Letter or S/U

Credit hours: 4

Recent offerings:

• http://www.cs.cornell.edu/courses/cs4702/2009sp
• http://www.cs.cornell.edu/courses/cs475/2006sp
  • Bart Selman