Hi there,

Unfortunately I have quite a backlog of other projects that I need to clear out, though Simple is AGPL so anyone is free to fork it. However I am also looking for employment opportunities.

Simple is already highly parallelizeable due to the fact that it relies on dynamic programming. Adding more threads of execution is just a matter of popping off another candidate test point from the priority queue while the objective function is already being evaluated at another point. I was planning to include this feature in a more mature release.

Any chance you could link the specific paper you’re talking about? This seems like a area I’d love to explore and the older you mentioned seems like a good starting point - I looked up the author on arxiv and there seems to be a few papers you could be referring to

This is the spearmint paper, describes how bayesian search algorithms can be parallelized in practice. There was also a very gentle introduction to bayesian optimization, I think it's one of the referenced papers. Cannot seem to find it right now.

https://arxiv.org/pdf/1206.2944.pdf