The CondorGP project came together over many years. Since 2005 in fact.
There are many pieces of the puzzle and my bet is that no-one else has cobbled together the bits I have, in the way that I have.
I should note there are some inspirational bits that I’m not going to talk about much (at all) as they are key to the design.
Self sustaining something
At one point at University, Alex, my coursemate and housemate and I were talking stuff. About how wouldn’t it be better if you could work on something, a business or an idea, and then when it was up and running then it would look after itself.
Seems minor, and obvious, but has influenced the thinking.
I came across Genetic Programming in an interesting textbook I bought at Brunel University, where Alex and I did an AI module as part of our engineering masters. This book was: Artificial Intelligence – A guide to intelligent systems, by Michael Negnevitsky
Alex and I worked together on an AI project, using ANNs, Artificial Neural Networks, like Deepmind used to create AlphaGo. AlphaGo came later, it must be said…
But flicking through again after University in 2005, I thought the sections on Genetic Programming were more interesting. I’ve still got the same copy of Negnevitsky’s book, so I type out directly:
“Genetic programming offers a solution to the main challenge of computer science – making computers solve problems without being explicitly programmed.”
This was pretty mind blowing and intriguing to a geeky engineering graduate. Further:
“Genetic programming represents an application of genetic model of learning to programming. Its goal is to evolve not a coded representation of some problem, but rather a computer code that solves the problem. That is, genetic programming generates computer programs as the solution.”
So the first steps of inspiration where that the power of genetic programming was enormous.
Key point 1 being that Darwin had hit upon the idea that evolution as a mechanism that had allowed or encouraged, or maybe actually been the basis for all the complexity and wonder of our planet. Key point 2 being that something of that power could be made available to us as part of computer work to solve complex problems. I quote again:
“evolutionary strategies can solve a wide range of problems. They provide robust and reliable solutions for highly complex, non-linear search and optimisation problems that previously could not be solved at all (Holland, 1995, Schwefel, 1995).”
So my brain immediately went to solving the obvious problem to a new graduate, how to make money. And directly to trading on the stock market.
I’ll look to write about genetic programming itself, in future posts, to share a little of how this works.