Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

UCB choice? #44

Closed
unhandyandy opened this issue Jul 17, 2018 · 5 comments
Closed

UCB choice? #44

unhandyandy opened this issue Jul 17, 2018 · 5 comments

Comments

@unhandyandy
Copy link

Have you thought about making the UCB a user definable property of the mctssolver type, as you did with init_Q?
@zsunberg
Copy link
Member

Yeah, see #14 . Would that be useful to you? If so, I can probably find time to implement it tomorrow.

@unhandyandy
Copy link
Author

No hurry, I've just over written best_sanode_UCB.

Also this by itself would useless without changing the StateNode type to include the state (my other issue yesterday). What I'm trying to do is implement the AlphaZero algorithm using Flux and MCTS. In that the UCB function depends on the neural net applied to the current state.

@zsunberg
Copy link
Member

Right, yeah it should be configurable and there needs to be a way to access the state.

@zsunberg zsunberg mentioned this issue Jul 26, 2018
Merged
@zsunberg
Copy link
Member

zsunberg commented Aug 9, 2018

the discussion for this is moving to #14

@zsunberg zsunberg closed this as completed Aug 9, 2018
@zsunberg
Copy link
Member

MCTS Criterion customization

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
2 participants
@unhandyandy @zsunberg