-
-
Notifications
You must be signed in to change notification settings - Fork 239
Feature/batch bo issue 1229 (QEI) #1236
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.
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
base: development
Are you sure you want to change the base?
Conversation
… feature/batch_bo_issue_1229
smac/main/config_selector.py
Outdated
| self._model, GaussianProcess | ||
| ), "Sample based estimate strategy only allows GP as surrogate model!" | ||
| return self._model.sample_functions(X_test=X_running, n_funcs=1) | ||
| elif estimation_strategy == "q_ei": |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Hmm... I think if we use q_ei, there is no need to estimate the running configurations. i.e., we use q_ei to provide multiple candidates at once.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I removed this part.
| if not self._log: | ||
| f_samples = normal_samples # in original (normal) space | ||
| f_min_sample = np.min(f_samples, axis=1) | ||
| improvement = np.maximum(self._eta - self._xi - f_min_sample, 0.0) |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Not quite sure here, but if the surrogate model is a GP, we can directly sample from the joint posterior distribution. As shown in botorch: https://botorch.readthedocs.io/en/stable/acquisition.html#botorch.acquisition.monte_carlo.qExpectedImprovement
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I replaced it with the surrogate model, if it is a GP.
I am not sure if it is correctly implemented. Please check!