I understand some recent chess engines (like alphazero or muzero) are based on neural networks. This question is not specific to chess, any other game (e.g. go) would do, but I keep chess for concreteness.

I am not interested how these engines are trained but how they "look like" after training. (For concreteness, it suffices for me that they give a "score" for a given board position.) Somewhat similar questions have been asked here but I can't find what I need.

  1. Is it correct to view them as a neural network? Or, is there some other major component like search trees or other type of algorithm. I'm assuming that answer here is "yes" in an approriate sense. Note: The alpha zero paper talks about various forms of tree search. It is not clear to me how they are used to play, and from the appendix apparently some other engines use very limited search, like depth 2? That would be already interesting to me.

  2. What is the depth of the network? Any reference to precise parameters?

  3. What is the input and output format of these networks? Is the input just an array with pieces positions (excluding three-fold repetition rules and similar secondary issues). What is the output? Some info is here but I can't quite understand how it works specifically e.g. for chess.

  4. What activation functions are used?

  • $\begingroup$ This is far too broad. $\endgroup$
    – Chenmunka
    Commented May 17 at 20:15
  • $\begingroup$ Please, vote to close it as "needs more focus" @Chenmunka $\endgroup$
    – nbro
    Commented May 18 at 13:47
  • $\begingroup$ Hello Manu. You're indeed asking too many questions in the same post. It's perfectly fine to have multiple questions and you questions seem interesting, but you should split this post into multiple ones, one for each question, unless they are really very related. As a rule of thumb, if you can't write your title as a specific question, then it means you need to split your post. $\endgroup$
    – nbro
    Commented May 18 at 13:48
  • $\begingroup$ OK. I put another question here ai.stackexchange.com/questions/45750/… $\endgroup$
    – Manu
    Commented May 19 at 14:00


You must log in to answer this question.

Browse other questions tagged .