According to a new Telegraph UK article, "a deep mathematical analysis of living things has come up with an insight into why they are special". The research was done by a researchers at Edinburgh University and the Max Planck Institute for Mathematics in the Sciences. What they've discovered is a way to make the behavior of artificial and virtual creatures self organizing and self regulating, allowing them to learn through trial and error without explicit instructions. The idea is not new and follows similar research going back to 1990. The article includes video of virtual dogs learning to jump and virtual humanoid learning to stand. Previous systems had explicitly encoded goals in the programs as well as basic instruction on how to move. The new algorithm, called homeokinesis, has no predefined goals, and provides no initial information about the virtual environment, making it more like the natural world. The researchers are also beginning to experiment with the algorithm in real robots. For more videos and papers on the new system, see Terra autonomica - The world of self-organized creatures.