Max Balandat
Max Balandat
Not only is using GPs + qEHVI slow in this regime, it's also not clear how you'd use the result effectively even if you could compute it faster? How are...
cc @ItsMrLin re the need for improved preference learning support in Ax :)
Is there observation noise in your setting? Winsorization is typically used as a way to handle outliers in the BO algorithm and less so in terms of performance evaluation (esp....
cc @sdaulton
I'd have to dig into the details what exactly is happening in Ax (@sdaulton would know better), but I believe it would be winsorization and then a standardization making both...
This seems like something that should be fixed in gpytorch (cc @gpleiss ) @lahramon your change was specifically to `get_fantasy_model`, have you investigated the effect of the state dict pre...
@Yoonho-Na can you provide a bit more detail on the specific problem you're running (how many parameters and what kind, is this single-objective or multi-objective, are there any constraints either...
This change make sense to me. It's technically BC breaking but it also fixes some consistency issue that's kind of a pain and has been open for a long time...
@hvarfner you have the best understanding of this - can you help out?
FWIW #1317 is an ancient issue that has some discussion on this :(. There is also a comment specifically on the `SmoothboxPrior` issue [here](https://github.com/cornellius-gp/gpytorch/pull/1737#discussion_r755628568).