Computing That Serves

General Imperfect Information Games


Thursday, March 1, 2018 - 11:00am


Mark Richards


Dan Ventura


Colloquium presented by Mark Richards
Thursday, March 1, 2018 at 11:00 A.M.
Location: 1170 TMCB

Several game-playing systems have been fine-tuned to outperform top human competition in individual disciplines such as chess, go, and some variants of poker. The next frontier is to develop a system that can generate AI for arbitrary games, given only a formal description of the rules. In this presentation, we will review results in general game playing for perfect information domains--where all agents have full knowledge of the game's history and current state--and present advances in representations, algorithms, and implementations for the general treatment of games with imperfect information.


Mark Richards earned B.S. and M.S. degrees in computer science from Brigham Young University and a Ph.D. in computer science from the University of Illinois at Urbana-Champaign. He has previously worked on surface-fitting methods at Sandia National Laboratories and on Bayesian inference techniques at Metron. Since 2012, he has been doing research and development of risk management software for banks at SAS Institute in Cary, North Carolina.