mlr3mbo
mlr3mbo copied to clipboard
mlr-org
Flexible Bayesian Optimization in R
reference = nadir + 1 is usually not appropriate if objectives live on vastly different scales.
````R require(data.table) require(mlr3mbo) require(bbotk) callback_plot = callback_optimization("plot", on_optimizer_after_eval = function(callback, context) { } ) obfun = ObjectiveRFun$new( fun = function(xs) 2 * xs$x * sin(14 * xs$x), domain = ps(x...
- [ ] input and output scaling of surrogates (input standardize, output normalize or log; retrafo or not) - [ ] adjust priors on length scales - [ ] kernels...
* check how delayed initialization allows for this * related to #19 * should also allow for constants in objective, see https://stackoverflow.com/questions/77500066/mlr3mbo-function-with-inputs-not-to-optimise
I.e., plot evaluated points, (surrogate prediction), acqf trace?
in general related to #110
The default initial design is evaluated in a single batch. if this batch evaluation takes long, the terminator cannot properly stop the process because it can only chime in after...
Hi team, Thanks for your great work on the collection of packages. I am a research in Australia using GPs to assist with deterministic models of forest dynamics. I am...
- [ ] include more examples on plots, e.g., using the surrogate
- [ ] docu - [ ] tests (more with respect to Nmin etc.) - [ ] handling of so vs. mo - [ ] wait for @be-marc to maybe...