1
$\begingroup$

I am coding a tic-tac-toe program that demonstrates reinforcement learning. The program uses minimax trees to decide its moves. Whenever it wins, all the nodes on the tree that were involved in the game have their value increased. Whenever it loses, all the nodes on the tree that were involved in the game have their value decreased, etc. What is the name of the value that each node is decreased by?

$\endgroup$
  • $\begingroup$ welcome to ai.stackexchange. Your question is a little bit too vague. Can you give us an example? For instance, I do not understand what you mean by increasing or decreasing all values which where involved in the tree. $\endgroup$ – marli Feb 1 '17 at 8:42
1
$\begingroup$

I think you mean the leave nodes only which are changed. The other nodes in the tree are calculated during calculating the best move with this tree. The values at the leaves are called utility values in Russel and Norvig's "Artificial intelligence: a modern approach". Some times it is called heuristic value; see https://en.wikipedia.org/wiki/Minimax.

$\endgroup$
0
$\begingroup$

In reinforcement learning, you can keep the name of value as it comes from the value function which estimates how good it is to be in this node w.r.t. the objective. Depending on the problem it can be a cost, utility, reward...

$\endgroup$

Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.