Non-Myopic Multifidelity Bayesian Optimization
This work addresses the challenge of accelerating optimization for expensive functions in fields like engineering or machine learning, but it appears incremental as it builds on existing multifidelity methods by extending the lookahead horizon.
The paper tackles the problem of optimizing black-box functions by proposing a non-myopic multifidelity Bayesian optimization framework that maximizes long-term rewards over two steps ahead, and it demonstrates that this algorithm outperforms standard multifidelity Bayesian methods on benchmark problems.
Bayesian optimization is a popular framework for the optimization of black box functions. Multifidelity methods allows to accelerate Bayesian optimization by exploiting low-fidelity representations of expensive objective functions. Popular multifidelity Bayesian strategies rely on sampling policies that account for the immediate reward obtained evaluating the objective function at a specific input, precluding greater informative gains that might be obtained looking ahead more steps. This paper proposes a non-myopic multifidelity Bayesian framework to grasp the long-term reward from future steps of the optimization. Our computational strategy comes with a two-step lookahead multifidelity acquisition function that maximizes the cumulative reward obtained measuring the improvement in the solution over two steps ahead. We demonstrate that the proposed algorithm outperforms a standard multifidelity Bayesian framework on popular benchmark optimization problems.