Robust Multi-Objective Bayesian Optimization Under Input Noise
Robust Multi-Objective Bayesian Optimization Under Input Noise
复制标题
DOI：
--
复制
发表时间：
2022-02
期刊：
影响因子：
--
通讯作者：
Sam Daulton;Sait Cakmak;M. Balandat;Michael A. Osborne;Enlu Zhou;E. Bakshy
中科院分区：
文献类型：
--
作者：
Sam Daulton;Sait Cakmak;M. Balandat;Michael A. Osborne;Enlu Zhou;E. Bakshy
贝叶斯优化(BO)是一种样本效率高的方法，用于调整设计参数，以优化评估昂贵的黑盒性能指标。在许多制造过程中，设计参数会受到随机输入噪声的影响，导致产品的性能往往不如预期。虽然已经提出了在输入噪声下优化单个目标的BO方法，但现有的方法还没有针对存在对输入扰动敏感的多个目标的实际场景。在这项工作中，我们提出了第一种对输入噪声具有鲁棒性的多目标BO方法。我们将我们的目标形式化为优化多变量风险价值(MVaR)，这是对不确定目标的风险度量。由于在许多情况下直接优化MVaR在计算上是不可行的，我们提出了一种基于随机标度的可扩展的、具有理论基础的MVaR优化方法。在实验中，我们发现我们的方法显著优于其他方法，并能有效地识别出最优的健壮设计，这些设计将以很高的概率满足多个指标的规格。
Bayesian optimization (BO) is a sample-efficient approach for tuning design parameters to optimize expensive-to-evaluate, black-box performance metrics. In many manufacturing processes, the design parameters are subject to random input noise, resulting in a product that is often less performant than expected. Although BO methods have been proposed for optimizing a single objective under input noise, no existing method addresses the practical scenario where there are multiple objectives that are sensitive to input perturbations. In this work, we propose the first multi-objective BO method that is robust to input noise. We formalize our goal as optimizing the multivariate value-at-risk (MVaR), a risk measure of the uncertain objectives. Since directly optimizing MVaR is computationally infeasible in many settings, we propose a scalable, theoretically-grounded approach for optimizing MVaR using random scalarizations. Empirically, we find that our approach significantly outperforms alternative methods and efficiently identifies optimal robust designs that will satisfy specifications across multiple metrics with high probability.