Programming computers using natural selection and genetic inheritance

Prof. Terence C. Fogarty

Since computers were invented about 50 years ago people have worked on getting them to programme themselves. During that time algorithms based on natural selection and genetic inheritance for the automatic production of computer programmes have been developed. Nevertheless, there is still an acute shortage of computer programmers and there is no likelihood of them being replaced by machines in the near future. So why can't we evolve computer programmes yet, what can we evolve artificially in a computer and will we ever get computers to evolve their own programmes?

Are evolutionary metaphors applicable to evolutionary optimization?

Dr. Andrew Tuson

Evolutionary (genetic) algorithms have proved to be effective optimisation methods for a wide range of difficult real-world problems. This success has lead to an assumption by some that evolutionary metaphors can form a basis for effective optimiser design. I will argue against this, describe some common fallacies that arise from this view, and introduce some alternative viewpoints.

The next step: Digital biology

Dr. Peter Bentley

There's a new breed of scientist emerging, called a digital biologist. These researchers perform interdisciplinary work, bringing together fields of biology with computer science, and enabling techniques such as neural networks and evolutionary computation to solve our problems more effectively. As we discover the limitations of these processes, we learn how natural biology moved beyond our feeble abilities millennia ago. We are now using immune systems, swarm intelligence, chaotic systems and embryology to extend the capabilities of neural networks and evolution. The technology of biology is far more advanced than any of our own approaches; digital biology harnesses these natural solutions.

Reference

This talk is loosely based on the content of the new popular science book Digital Biology by Peter J Bentley, to be published early next year.