-
Notifications
You must be signed in to change notification settings - Fork 1
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
Incremental bayesian optimisation? #144
Comments
Hi. Could you elaborate? Do you have some suggestions or paper references in mind? At the moment I have just made a straightforward implementation without retaining any state. |
Here is the test for #[test]
fn test_xsinx_egor_builder() {
let ego = EgorServiceBuilder::optimize()
.configure(|conf| {
conf.regression_spec(RegressionSpec::ALL)
.correlation_spec(CorrelationSpec::ALL)
.infill_strategy(InfillStrategy::EI)
.random_seed(42)
})
.min_within(&array![[0., 25.]]);
let mut doe = array![[0.], [7.], [20.], [25.]];
let mut y_doe = xsinx(&doe.view());
for _i in 0..10 {
let x_suggested = ego.suggest(&doe, &y_doe);
doe = concatenate![Axis(0), doe, x_suggested];
y_doe = xsinx(&doe.view());
}
let expected = -15.1;
let y_opt = y_doe.min().unwrap();
assert_abs_diff_eq!(expected, *y_opt, epsilon = 1e-1);
} You have Does this imply that for each iteration there is This crate: https://crates.io/crates/friedrich claims that they are able to process additional samples in I want to integrate |
Yes each iteration implies a Otherwise I forgot about this feature in |
In any MCMC-based heuristic, such as simulated annealing, parallel tempering or anything else, the number of steps is very large. We are talking about perhaps up to hundreds of thousands or millions. For this, it would be very helpful if |
I've recently implemented sparse GP methods in the |
I want to use 5700 comes from the number of qubits that the biggest quantum annealers currently have. I wrote https://github.com/brurucy/ernst to semi-simulate that. |
Do you mean beside having n ~ 1e5 samples, your input dimension is 5700?! 😮 |
Up to that, yes. |
In |
There's a caveat. All of these variables are binary: +1 or -1. |
Closing stale issue, feel free to open a fresh one if needed. |
Hi,
I wonder, is it possible for the service struct to incrementally adjust to new suggestions, instead of recomputing everything from scratch in every iteration?
The text was updated successfully, but these errors were encountered: