BINOCULARS for Efficient, Nonmyopic Sequential Experimental Design

Abstract

Finite-horizon sequential experimental design (SED) arises naturally in many contexts, including hyperparameter tuning in machine learning among more traditional settings. Computing the optimal policy for such problems requires solving Bellman equations, which are generally intractable. Most existing work resorts to severely myopic approximations by limiting the decision horizon to only a single time-step, which can underweight exploration in favor of exploitation. We present BINOCULARS: Batch-Informed NOnmyopic Choices, Using Long-horizons for Adaptive, Rapid SED, a general framework for deriving efficient, nonmyopic approximations to the optimal experimental policy. Our key idea is simple and surprisingly effective: we first compute a one-step optimal batch of experiments, then select a single point from this batch to evaluate. We realize BINOCULARS for Bayesian optimization and Bayesian quadrature -- two notable SED problems with radically different objectives -- and demonstrate that BINOCULARS significantly outperforms myopic alternatives in real-world scenarios.

arXiv comment13 pages, 4 figures, 6 tables, 1 algorithm

References in corpus (10)

• Practical Bayesian Optimization of Machine Learning Algorithms
• Batch Bayesian Optimization via Local Penalization
• Max-value Entropy Search for Efficient Bayesian Optimization
• The Parallel Knowledge Gradient Method for Batch Bayesian Optimization
• Parallel Bayesian Global Optimization of Expensive Functions
• Parallel Predictive Entropy Search for Batch Global Optimization of Expensive Objective Functions
• GLASSES: Relieving The Myopia Of Bayesian Optimisation
• Sampling for Inference in Probabilistic Models with Fast Bayesian Quadrature
• Improving Quadrature for Constrained Integrands
• Automated Model Selection with Bayesian Quadrature