Ponente: Bob McKay (Seoul National University)

Fecha: martes 6 de julio 2010

Hora: 11h00

Lugar de celebración: Salón de Grados, ETSI Industriales, UNED

Grammars in Genetic Programming Estimation of Distribution Algorithms: Why and How? [slides]

Resumen

Grammars have been widely used in Genetic Programming as a means to describe the search space, but are still a small proportion of overall GP work. In the smaller field, of GP systems using Estimation of Distribution Algorithms, they have come to dominate the field. In this talk, we explain why, and introduce the interesting grammar learning problems that they give rise to. We introduce the problem domain of Genetic Programming, and the general technique of Estimation of Distribution Algorithms, and then briefly discuss the representations used, and why (we believe) grammar-based representations have come to dominate. We outline the different kinds of stochastic grammar models that have been used, and the grammar learning methods that have been applied. We then attempt to outline the intrinsic grammar-learning issues that arise, with the hope of stimulating discussion of options for grammar learning in EDA-GP.

Lugar de Celebración

Cómo llegar y planos