Bayesian Optimization for Min Max Optimization

A solution that is only reliable under favourable conditions is hardly a safe solution. Min Max Optimization is an approach that returns optima that are robust against worst case conditions. We propose algorithms that perform Min Max Optimization in a setting where the function that should be optimized is not known a priori and hence has to be learned by experiments. Therefore we extend the Bayesian Optimization setting, which is tailored to maximization problems, to Min Max Optimization problems. While related work extends the two acquisition functions Expected Improvement and Gaussian Process Upper Confidence Bound; we extend the two acquisition functions Entropy Search and Knowledge Gradient. These acquisition functions are able to gain knowledge about the optimum instead of just looking for points that are supposed to be optimal. In our evaluation we show that these acquisition functions allow for better solutions - converging faster to the optimum than the benchmark settings.

• Publication year

  2021

• Venue

  arXiv.org

• Publication date

  2021-07-29

• Fields of study

  Mathematics, Computer Science

• Identifiers
  arXiv 2107.13772
• External record

  Open on Semantic Scholar

• Source metadata

  Semantic Scholar

• No claims are published for this paper.

• No concepts are published for this paper.

• Max-value Entropy Search for Multi-Objective Bayesian Optimization

  2020cited by this paper
• Mixed Strategies for Robust Optimization of Unknown Objectives

  2020cited by this paper
• Noisy-Input Entropy Search for Efficient Robust Bayesian Optimization

  2020influential reference
• BoTorch: Programmable Bayesian Optimization in PyTorch

  2019cited by this paper
• Integrals over Gaussians under Linear Domain Constraints

  2019cited by this paper
• Bayesian Optimization with Expensive Integrands

  2018cited by this paper
• A Tutorial on Bayesian Optimization

  2018influential reference
• Adversarially Robust Optimization with Gaussian Processes

  2018cited by this paper
• Max-value Entropy Search for Efficient Bayesian Optimization

  2017cited by this paper
• Taking the Human Out of the Loop: A Review of Bayesian Optimization

  2016cited by this paper
• A new expected-improvement algorithm for continuous minimax optimization

  2015cited by this paper
• Efficient global robust optimization of unconstrained problems affected by parametric uncertainties

  2015cited by this paper
• Automatic testing and minimax optimization of system parameters for best worst-case performance

  2015influential reference
• Entropy Search for Information-Efficient Global Optimization

  2011cited by this paper
• Sparse Spectrum Gaussian Process Regression

  2010cited by this paper
• The Knowledge-Gradient Policy for Correlated Normal Beliefs

  2009cited by this paper
• Expectation-Propogation for the Generative Aspect Model

  2002cited by this paper
• Efficient Global Optimization of Expensive Black-Box Functions

  1998influential reference
• ON THE LIKELIHOOD THAT ONE UNKNOWN PROBABILITY EXCEEDS ANOTHER IN VIEW OF THE EVIDENCE OF TWO SAMPLES

  1933influential reference

• Asymmetrical evolutionary strategy with adaptive Bayesian optimization for robust design of complex electromechanical product

  2026cites this paper
• Bayesian Optimization for Function-Valued Responses under Min-Max Criteria

  2025cites this paper
• Robust Entropy Search for Safe Efficient Bayesian Optimization

  2024cites this paper
• Spectral Representation of Robustness Measures for Optimization Under Input Uncertainty

  2022cites this paper