Hopefully this post will grow over the next few days as I have time to recap the past month. In short here are a few notable events in the life of this particular White…
- I felt an earthquake first hand. Small, but still my first earthquake
- We had a freak 15 minute ice storm followed by a weekend of snow (in April)
- I just recieved word that QOpenOCCI may be included in KDE if I so choose
- I got to go to Oaklawn, though it was not nearly as profitable as past trips
- My AI project is do well and is again a magnitude faster due to some clever coding tricks
- The Merlin got an echo
- Got some new pictures uploaded from Oaklawn and of the pets.
I wrote an article earlier that discussed AI and learning. Recently I have been working on a new prototype that further expands on some of my ideas. I call it the Generic Probabilistic Artificial Neural Network. Though the first version is really slow, it was able to evolve to the point of solving the XOR function.
I would like to note that the AI evolved to this point and that it did not learn the XOR function. That is, it used a base program that did nothing and mutated the copies of that program and kept the most fit for further mutation.
My results showed that on average after 125 generations with a 10x reproduction rate and a 1 step mutation rate, the XOR solution was found and the program was between 100-300 neurons in size.
I am currently working on optimizing how the simulation rus to better increase performance. I also wish to dive into learning soon. That is, the AI evolves to solve arbitrary problems using feedback and memory.