I am a student new in reinforcement learning and I'm trying to implement an AI able to play Checkers. I want to implement a deep learning solution.

However, I am confused on how to do that. I understood the concept of states, actions, and rewards, but when I read research papers, generally, the actions are always the same independently to the state, like moving up or down or left or right. In checkers, it's a bit more complicated because you can't know in advance the actions you can or cannot do, you might be able to jump over a checker, go back if you're a king, etc, etc and it all depends on the state of each checkers (alive, king, dead).

For now, I have implemented the Board, which would be the environment, but I'm stuck on what to feed to the neural network and what to expect as output.

So my question would be, how is it possible to implement a Deep reinforcement learning algorithm when you have the actions that change depending on the state?

Thanks in advance.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.