Hai everyone,

We are implementing a self-learning chess player using neural networks and other concepts like genetic algos.

We have implemented the neural network module. There are 4 neural networks representing 16 squares of the chessboard each. Thus the material values of a chess peice on a square act as an input value(positive for friendly peices and negative for opponent peices).

Thus each neural network has 16 input nodes and one output giving us the strategic position of the player in that region.

All we need is the training sets for the neural networks. Can anyone tell us where we can acquire the training sets that correspond to our project? Or is there a tool that can generate the training sets for our chess project?

Our project cannot continue without training the neural networks. So we would really appreciate any help. Thanks in advance. :)

m*a*s team

If this is any help, our neural networks each have 3 layers of hidden neurons and also have error back propagation that automatically adjust weights. All the weights are stored in binary files.

If someone can provide the training set in text/binary format, we can easily integrate it into our project. But any help to find training set is also gr8.

m*a*s team