ICML-98 Submission #137

Coevolutionary Learning: A Case Study

Author: Hugues Juille and Jordan B. Pollack
	Computer Science Department
	Brandeis University
	Waltham, Massachusetts 02254-9110

Abstract:

Coevolutionary learning, which involves the embedding of adaptive
learning agents in a fitness environment that dynamically responds to
their progress, is a potential solution for many technological chicken
and egg problems and is at the heart of some recent and surprising
successes such as Tesauro's backgammon player \cite{Tesauro95}.
However, several impediments have to be overcome in order for
coevolutionary learning to achieve continuous progress in the long
term.  This paper presents some of those problems and proposes a
framework to address them.  This presentation is illustrated with a
case study: the evolution of CA rules.  Our application of
coevolutionary learning resulted in a very significant improvement for
that problem compared to the best known results.

Keywords: Adaptive environment, Coevolution, Self-adaptation, Cellular Automata.

Contact author: Hugues Juille
	E-mail: hugues@cs.brandeis.edu
	Phone:  (781) 736-2743