After implementing the predictor in the main DM program, I had the chance to run the and them dump percepts to give some form to the ML generated output previously posted. The results were quite weak. It appears that there is simply too much information to be encapsulated by the small number (~2000) of percepts […]
↧