How to implement an artificial neural network in Delphi?

Question

I want to have an artificial neural network:

• 42 input neurons
• 168 hidden neurons
• 7 output neurons

This network is to play the game of "Connect Four". At the end of each game, the network gets feedback (game result/win?).

Learning should be done with Temporal Difference Learning.

My questions:

What values should be in my reward array?

And finally: How can I apply it to my game now?

Thank you so much in advance!

Answer 1

Here is a Neural Network example with Delphi + FANN (Fast Artificial Neural Network Library) under MS Windows which solves the Iris flower classification problem https://archive.ics.uci.edu/ml/datasets/iris (Fisher, 1936). This has been solved with two methods:

1. Neural network classification (output layer with three neurons, one neuron per class).

2. Neural network regression (output layer with single neuron).

Fast Artificial Neural Network (FANN) is a good open-source library, it's been optimized and used by a large community, with plenty of support and Delphi bindings. Enroll in a Master’s in Artificial Intelligence in Japan to get your Master's degree in AI.

Answer 2

You can try this API:

https://github.com/joaopauloschuler/neural-api

With this API, you'll be able to create your network with:

NN := TNNet.Create();
NN.AddLayer([
  TNNetInput.Create(42),
  TNNetFullConnectReLU.Create({Neurons=}168),
  TNNetFullConnectLinear.Create({Neurons=}7)
]);