I spent six weeks in Spain this summer. My wife, Tracy, is earning a Master’s Degree in
Spanish from St. Louis University (SLU). Her university has an extension campus in Madrid.
I also wanted to be in Spain to visit the GECCO-2015 conference, that was held in Madrid.
We were able to line up a nice apartment in Salamanca, Madrid and I was able to take a
few weeks vacation and work remotely the rest of the time. It was totally worth it, Spain
is an amazing country. I know just enough Spanish to get around and usually be fed at a
The GECCO conference was awesome. I was able to briefly meet Dr. Kenneth Stanley after
hearing a tutorial session held by him. I’ve many of Dr. Stanley’s papers and implemented
his NEAT/HyperNEAT algorithms. He also has a new book out that I’ve just read and recommend.
The book points out that it is very difficult to design objective functions that might
truly create amazing results. Some of the best results on Ken’s PicBreader site were not
I also saw a fascinating Lisp-based language named Push3, created by Dr. Lee Spector.
Genetic programming ha always been fascinating to me because it uses AI to actually
evolve AI. Push3 is a programming language designed for computers to evolve.
Dr. Spector’s group is evolving computer programs, to accomplish basic tasks like word
count and other basic utilities. Most machine learning algorithms are simply about
optimizing coefficients to decrease a supervised training set’s error. It is amazing to
see algorithms that can actually evolve themselves into new algorithms.
I plan to make use of genetic programming for part of my dissertation. Of course, I am
still finishing up coursework and this is very much in flux.