Open ghutchis opened 1 year ago
botorch expected improvement methods don't give the $\epsilon$ hyper parameter to push for more exploration.
It might be worth modifying the "best value" (e.g., best + 0.01) to push for more exploration like with the current UCB method.
For now, the code uses UCB with a formula favoring increasing exploration, which might make sense for finding all minima.
Benchmark # of steps required and lowest minima across a subset of the BOKEI test set:
Other interesting acquisition functions in BoTorch: