Max-value Entropy Search for Efficient Bayesian Optimization

Abstract

Entropy Search (ES) and Predictive Entropy Search (PES) are popular and empirically successful Bayesian Optimization techniques. Both rely on a compelling information-theoretic motivation, and maximize the information gained about the $\arg \max$ of the unknown function; yet, both are plagued by the expensive computation for estimating entropies. We propose a new criterion, Max-value Entropy Search (MES), that instead uses the information about the maximum function value. We show relations of MES to other Bayesian optimization methods, and establish a regret bound. We observe that MES maintains or improves the good empirical performance of ES/PES, while tremendously lightening the computational burden. In particular, MES is much more robust to the number of samples used for computing the entropy, and hence more efficient for higher dimensional problems.

Proceedings of the 34th International Conference on Machine Learning, Sydney, Australia, PMLR 70, 2017

Cited by in corpus (92)

• An Efficient Batch Constrained Bayesian Optimization Approach for Analog Circuit Synthesis via Multi-objective Acquisition Ensemble
• Greed is Good: Exploration and Exploitation Trade-offs in Bayesian Optimisation
• Bayesian Optimization with Output-Weighted Optimal Sampling
• Upper Trust Bound Feasibility Criterion for Mixed Constrained Bayesian Optimization with Application to Aircraft Design
• Adaptive and Safe Bayesian Optimization in High Dimensions via One-Dimensional Subspaces
• Efficient acquisition rules for model-based approximate Bayesian computation
• Cost-aware Bayesian Optimization
• Distributionally Robust Bayesian Optimization
• Multi-fidelity Bayesian Optimization with Max-value Entropy Search and its parallelization
• Multi-Fidelity Bayesian Optimization via Deep Neural Networks
• Noisy-Input Entropy Search for Efficient Robust Bayesian Optimization
• Constrained Bayesian Optimization with Max-Value Entropy Search
• Zeroth-Order Algorithms for Nonconvex Minimax Problems with Improved Complexities
• Local policy search with Bayesian optimization
• What do you Mean? The Role of the Mean Function in Bayesian Optimisation
• A Probabilistic Machine Learning Approach to Scheduling Parallel Loops with Bayesian Optimization
• Constant-Time Predictive Distributions for Gaussian Processes
• Are we Forgetting about Compositional Optimisers in Bayesian Optimisation?
• Output Space Entropy Search Framework for Multi-Objective Bayesian Optimization
• Multi-objective Bayesian Optimization using Pareto-frontier Entropy
• Time Efficiency in Optimization with a Bayesian-Evolutionary Algorithm
• Leveraging Trust for Joint Multi-Objective and Multi-Fidelity Optimization
• Max-value Entropy Search for Multi-Objective Bayesian Optimization with Constraints
• Fast Matrix Square Roots with Applications to Gaussian Processes and Bayesian Optimization
• Knowing The What But Not The Where in Bayesian Optimization
• Adaptive and Collaborative Bathymetric Channel-Finding Approach for Multiple Autonomous Marine Vehicles
• Epsilon-Greedy Thompson Sampling to Bayesian Optimization
• Speeding Up Multi-Objective Hyperparameter Optimization by Task Similarity-Based Meta-Learning for the Tree-Structured Parzen Estimator
• Pareto Front-Diverse Batch Multi-Objective Bayesian Optimization
• Batch simulations and uncertainty quantification in Gaussian process surrogate approximate Bayesian computation
• Scalable Combinatorial Bayesian Optimization with Tractable Statistical models
• GIBBON: General-purpose Information-Based Bayesian OptimisatioN
• Corruption-Tolerant Gaussian Process Bandit Optimization
• On Lower Bounds for Standard and Robust Gaussian Process Bandit Optimization
• Multi-objective Asynchronous Successive Halving
• How Bayesian Should Bayesian Optimisation Be?
• Deep Multi-Fidelity Active Learning of High-dimensional Outputs
• Excursion Search for Constrained Bayesian Optimization under a Limited Budget of Failures
• A New Knowledge Gradient-based Method for Constrained Bayesian Optimization
• Adaptive Expansion Bayesian Optimization for Unbounded Global Optimization
• A Domain-Shrinking based Bayesian Optimization Algorithm with Order-Optimal Regret Performance
• Information-Theoretic Multi-Objective Bayesian Optimization with Continuous Approximations
• Mono-surrogate vs Multi-surrogate in Multi-objective Bayesian Optimisation
• Optimal Order Simple Regret for Gaussian Process Bandits
• Cautious Bayesian Optimization for Efficient and Scalable Policy Search
• Pareto-efficient Acquisition Functions for Cost-Aware Bayesian Optimization
• Multipoint-BAX: A New Approach for Efficiently Tuning Particle Accelerator Emittance via Virtual Objectives
• Incorporating Expert Prior Knowledge into Experimental Design via Posterior Sampling
• Multi-Step Budgeted Bayesian Optimization with Unknown Evaluation Costs
• Gradient Free Minimax Optimization: Variance Reduction and Faster Convergence
• An Asymptotically Optimal Multi-Armed Bandit Algorithm and Hyperparameter Optimization
• Towards Automatic Bayesian Optimization: A first step involving acquisition functions
• Bayesian Algorithm Execution: Estimating Computable Properties of Black-box Functions Using Mutual Information
• BOSH: Bayesian Optimization by Sampling Hierarchically
• BoA-PTA, A Bayesian Optimization Accelerated Error-Free SPICE Solver
• Quantity vs. Quality: On Hyperparameter Optimization for Deep Reinforcement Learning
• A portfolio approach to massively parallel Bayesian optimization
• No-Regret Bayesian Optimization with Unknown Hyperparameters
• Efficient Nonmyopic Bayesian Optimization via One-Shot Multi-Step Trees
• Value-at-Risk Optimization with Gaussian Processes
• Bayesian Optimization for Iterative Learning
• BINOCULARS for Efficient, Nonmyopic Sequential Experimental Design
• Optimizing Photonic Nanostructures via Multi-fidelity Gaussian Processes
• Practical Transfer Learning for Bayesian Optimization
• Reinforced Few-Shot Acquisition Function Learning for Bayesian Optimization
• Sequential- and Parallel- Constrained Max-value Entropy Search via Information Lower Bound
• Efficient Batch Black-box Optimization with Deterministic Regret Bounds
• Efficient Transfer Bayesian Optimization with Auxiliary Information
• Using Distance Correlation for Efficient Bayesian Optimization
• Lenient Regret and Good-Action Identification in Gaussian Process Bandits
• Batch Multi-Fidelity Bayesian Optimization with Deep Auto-Regressive Networks
• BONSAI: Structure-exploiting robust Bayesian optimization for networked black-box systems under uncertainty
• BORE: Bayesian Optimization by Density-Ratio Estimation
• Mixed Strategies for Robust Optimization of Unknown Objectives
• Bayesian Optimization using Pseudo-Points
• Classified Regression for Bayesian Optimization: Robot Learning with Unknown Penalties
• Multi-Fidelity Multi-Objective Bayesian Optimization: An Output Space Entropy Search Approach
• Multi-objective Bayesian optimization of ferroelectric materials with interfacial control for memory and energy storage applications
• Adversarial Attacks on Gaussian Process Bandits
• Gaussian Process Sampling and Optimization with Approximate Upper and Lower Bounds
• Recursive Two-Step Lookahead Expected Payoff for Time-Dependent Bayesian Optimization
• MBORE: Multi-objective Bayesian Optimisation by Density-Ratio Estimation
• RTDK-BO: High Dimensional Bayesian Optimization with Reinforced Transformer Deep kernels
• Lookahead Acquisition Functions for Finite-Horizon Time-Dependent Bayesian Optimization and Application to Quantum Optimal Control
• Practical Bayesian Optimization of Objectives with Conditioning Variables
• Trusted-Maximizers Entropy Search for Efficient Bayesian Optimization
• Efficient Bayesian Optimization using Multiscale Graph Correlation
• One-parameter family of acquisition functions for efficient global optimization
• Robust Policy Search for Robot Navigation
• OpEvo: An Evolutionary Method for Tensor Operator Optimization
• Bayesian optimization for modular black-box systems with switching costs
• Beyond Marginal Uncertainty: How Accurately can Bayesian Regression Models Estimate Posterior Predictive Correlations?