A work-in-progress to use Neural Networks to recognize expert level moves in the game of Go. So far the best network produced by this code has an accuracy rate of 3.5%, which should be drastically improved upon as the project continues.

The games used as training data for this project come from the GoGoD Winter 2010 edition. For copyright as well as size reasons, they can not be distributed with this project.

Instructions on running this project to follow as the project matures.