Visiting Fellowship: Bringing contemporary biology into Evolutionary Computation: Plasticity, hierarchy, and genetic re-use

点击翻译按钮获取中文摘要

Evolutionary computation is a branch of computer science using ideas from Darwinian evolution and genetics to design systems and algorithms. Starting from a primordial ooze of random tentative solutions to a problem, generation after generation, evolutionary algorithms refine them using genetic mutations and recombinations and the principle of survival of the fittest. This ultimately leads to the creation of highly fit novel solutions to the problem at hand. Evolutionary computation has provided tens of remarkable human-competitive results. However, there are clear limits in terms of the size and complexity of the kinds of solutions we are able to currently evolve. The main problems are: * limited modularity (modularity is the ability to build large systems out of smaller, semi-independent ones), * limited scalability (scalability refers to how rapidly the computational effort of running a problem-solver increases with the size of the problems),* lack of plasticity (plasticity refers to the ability for a solution to change its behaviour in response to external conditions and events)This project will tackle all three of these significant problems in evolutionary computation with a radically new approach, inspired by biology that has already solved these problems.