till-m

Hey @Rizhiy, thanks for contributing! I (superficially) follwed the discussion in the original issue and I agree with the notions that a) this package should be compatible with parallel optimization...

Hi @Rizhiy, I misread the code and after thinking about it again, I think you've covered the scenario that I was picturing. My apologies.

Hi @Rizhiy, > I would guess to make it more user-friendly, the relevant functions (e.g. max & res) should include the filter for dummies. I agree. I think logging should...

Hi @Rizhiy, I decided to write a quick implementation of Kandasamy to see it work. The preliminary results are pretty disappointing, will post later.

Okay, so first of all it looks like sampling from the multivariate normal is _very_ computationally expensive. I had to adapt `n_warmup` in `acq_max` to make this feasible, so I'm...

My apologies for the delay. > Is there something similar to kappa decay in your code? In theory this shouldn't be necessary, but I could add it. I will try...

Hi @Rizhiy, apologies for letting this sit for a while. I've been thinking and we should definitely have functionality for this, but it would be better to first redesign the...

Closing as this is handled in #447. @Rizhiy I adapted your benchmarking code as an [example](https://github.com/bayesian-optimization/BayesianOptimization/blob/master/examples/async_optimization_dummies.py) incl. attribution. I hope that's okay.

Could you maybe check if this was fixed with #437? I think we haven't made a release since that PR, but you can install from master here. E: If I...

Something like this has also been raised in #381. I'm still of the opinion I expressed there, i.e. that `.maximize` should be kept simple and that we should leave these...