Are we Forgetting about Compositional Optimisers in Bayesian Optimisation?