The Math and Statistics Club Presents:

What has evolved? Training game playing agents with evolutionary algorithms.

Speaker: Daniel Ashlock,
Credentials: Professor of Mathematics, Bioinformatics Chair, Senior Member IEEE, University of Guelph
Time and Place: 5:30 pm - Wednesday, October 20 in 1511 SCIE

Snacks Provided.

Abstract: It has been twenty years since John Miller first trained finite state machines to play the game iterated prisoner's dilemma with a digital version of Darwin's theory of evolution. This talk will introduce the area of evolutionary game theory and cover several recent advances including a technique called fingerprinting that permits automatic analysis of game playing agents and recent work on endowing computer game-playing agents with emotion. Emotional agents are found to have a substantial competitive advantage. The talk is intended for a general audience.