Unpublished,
Machine Learning with Genetic Programming: Some Challenges
(November 2003)Keynote Address.
Abstract
Genetic Programming has achieved a great deal in the
twenty or so years since its inception. Nevertheless,
there remain a wide range of challenges and open
questions: what form do the building blocks of good
programs take; how may they be identified and promoted?
Particularly in the light of Daida's work on the
structural restrictions of GP, there are critical
questions on the appropriate representations for GP,
and the possibility of non-traditional genetic
operators. Finally, it is clear that the operation of
GP systems still differs in important ways from that of
biological systems, so there is opportunity for further
inspiration from biological systems.
Tags
- algorithms,
- bias,
- developmental,
- genetic
- genotype-phenotype
- hierarchical,
- mapping
- programming,
- search
