Predictive Entropy Search for Efficient Global Optimization of Black-box Functions

Abstract

We propose a novel information-theoretic approach for Bayesian optimization called Predictive Entropy Search (PES). At each iteration, PES selects the next evaluation point that maximizes the expected information gained with respect to the global maximum. PES codifies this intractable acquisition function in terms of the expected reduction in the differential entropy of the predictive distribution. This reformulation allows PES to obtain approximations that are both more accurate and efficient than other alternatives such as Entropy Search (ES). Furthermore, PES can easily perform a fully Bayesian treatment of the model hyperparameters while ES cannot. We evaluate PES in both synthetic and real-world applications, including optimization problems in machine learning, finance, biotechnology, and robotics. We show that the increased accuracy of PES leads to significant gains in optimization performance.

References in corpus (3)

• Practical Bayesian Optimization of Machine Learning Algorithms
• A Tutorial on Bayesian Optimization of Expensive Cost Functions, with Application to Active User Modeling and Hierarchical Reinforcement Learning
• Adaptive Hamiltonian and Riemann Manifold Monte Carlo Samplers

Cited by in corpus (12)

• Predictive Entropy Search for Bayesian Optimization with Unknown Constraints
• Multi-fidelity Bayesian Optimisation with Continuous Approximations
• Parallel and Distributed Thompson Sampling for Large-scale Accelerated Exploration of Chemical Space
• An Entropy Search Portfolio for Bayesian Optimization
• Batched Gaussian Process Bandit Optimization via Determinantal Point Processes
• Actively Learning what makes a Discrete Sequence Valid
• Query-limited Black-box Attacks to Classifiers
• Towards Information-Seeking Agents
• PHOENICS: A universal deep Bayesian optimizer
• Advancing Bayesian Optimization: The Mixed-Global-Local (MGL) Kernel and Length-Scale Cool Down
• Discretization-free Knowledge Gradient Methods for Bayesian Optimization
• Bayesian Optimization for Probabilistic Programs