EnviGP - Improving Genetic Programming for the Environment and Other Applications

Description

Genetic Programming (GP) is the youngest paradigm inside the artificial intelligence research area called evolutionary computation, and consists on the automated learning of computer programs. GP often yields results that are not merely academically interesting, but competitive with the work developed by humans. However, because it is a young and complex paradigm, the practical use of GP still poses a few challenges. In this project we will develop and test new approaches to the bloat and overfitting problems in GP, while studying the relationship between the two, and adapt GP for improved efficiency in multiclass classification problems. The achievement of these goals will ultimately produce a powerful general-purpose tool that can be used by practitioners of many diverse areas of research.

Researchers

Funded by

FCT

Partners

INESC-ID, FCTUC, IICT, University of Milano-Bicocca

Start Date

2010-04-16

End Date

2013-10-15

Journal Articles

Conference Articles

2013

(3 publications)

2012

(1 publication)

2011

(6 publications)

2010

(6 publications)

Book Chapters