Asynchronous ε-Greedy Bayesian Optimisation
This work addresses the need for efficient optimization in machine learning and engineering by reducing wallclock time through asynchronous evaluation, though it is incremental as it builds on existing Bayesian optimization techniques.
The paper tackles the problem of optimizing expensive black-box functions by developing AEGiS, an asynchronous Bayesian optimization method that combines greedy search, Thompson sampling, and random selection to maximize resource utilization, and demonstrates that it generally outperforms existing asynchronous methods while matching performance with a single worker.
Batch Bayesian optimisation (BO) is a successful technique for the optimisation of expensive black-box functions. Asynchronous BO can reduce wallclock time by starting a new evaluation as soon as another finishes, thus maximising resource utilisation. To maximise resource allocation, we develop a novel asynchronous BO method, AEGiS (Asynchronous $ε$-Greedy Global Search) that combines greedy search, exploiting the surrogate's mean prediction, with Thompson sampling and random selection from the approximate Pareto set describing the trade-off between exploitation (surrogate mean prediction) and exploration (surrogate posterior variance). We demonstrate empirically the efficacy of AEGiS on synthetic benchmark problems, meta-surrogate hyperparameter tuning problems and real-world problems, showing that AEGiS generally outperforms existing methods for asynchronous BO. When a single worker is available performance is no worse than BO using expected improvement.