The CGP web site has been moved to here and the domain name cartesiangp.co.uk no longer exists.
Julian is giving a tutorial on CGP at Alife 2019. He is also co-organizing a second workshop on Developmental Neural Networks. We are looking for sponsors to fund an exciting invited speaker. Developmental ANNs are concerned with using ideas inspired by the development of brains to create programs that build neural networks.
Human-competitive results workshop (GECCO 2018)
CGP was used by two of the awarded entries this year. One won first prize, the other bronze. Great stuff.
The CGP paper on atari games also attracted lots of media attention (see publications).
What is Julian upto?
CGP Brain project
For a couple of years now I have been working on using CGP to evolve two programs that develop a whole neural network and can learn to solve multiple problems simultaneously. One evolved program controls the body of a neuron and the other controls the dendrites. When you run the programs it build an entire brain-like network which can solve three problems simultaneously (classification). The aim is to solve many problems of different types with the same network and the evolved programs. You can read a brief conference report on it (GECCO 2017). I have two book chapters coming out on this topic (see publications) which go into a lot of detail.
The current status and future of CGP
I have recently submitted a paper on CGP: its status and future to the 20th anniversary edition of the journal Genetic Programming and Evolvable Machines. It reviews many aspects of CGP and discusses open issues and new ideas. I am hoping the reviewers will like it and it will get published. It discusses so many things that can make CGP even better... :-)
I have also recently submitted a journal article on what I call computational alchemy. I give two examples: 1. CGP and 2. Evolution-in-materio. In the former strange interactions between software components can be exploited by evolutionary algorithms and in the latter strange interactions between physical componets are exploited. Hopefully, the article will appear in the journal Natural Computing at some point.