Parallel and Distributed Thompson Sampling for Large-scale Accelerated Exploration of Chemical Space

Abstract

Chemical space is so large that brute force searches for new interesting molecules are infeasible. High-throughput virtual screening via computer cluster simulations can speed up the discovery process by collecting very large amounts of data in parallel, e.g., up to hundreds or thousands of parallel measurements. Bayesian optimization (BO) can produce additional acceleration by sequentially identifying the most useful simulations or experiments to be performed next. However, current BO methods cannot scale to the large numbers of parallel measurements and the massive libraries of molecules currently used in high-throughput screening. Here, we propose a scalable solution based on a parallel and distributed implementation of Thompson sampling (PDTS). We show that, in small scale problems, PDTS performs similarly as parallel expected improvement (EI), a batch version of the most widely used BO heuristic. Additionally, in settings where parallel EI does not scale, PDTS outperforms other scalable baselines such as a greedy search, $\epsilon$-greedy approaches and a random search method. These results show that PDTS is a successful solution for large-scale parallel BO.

Accepted for publication in the proceedings of the 2017 ICML conference

References in corpus (7)

• Weight Uncertainty in Neural Networks
• Probabilistic Backpropagation for Scalable Learning of Bayesian Neural Networks
• Massively Multitask Networks for Drug Discovery
• Predictive Entropy Search for Efficient Global Optimization of Black-box Functions
• Batch Bayesian Optimization via Local Penalization
• Parallel Bayesian Global Optimization of Expensive Functions
• An Entropy Search Portfolio for Bayesian Optimization

Cited by in corpus (22)

• Less is more: sampling chemical space with active learning
• Accelerating high-throughput virtual screening through molecular pool-based active learning
• Gryffin: An algorithm for Bayesian optimization of categorical variables informed by expert knowledge
• Automated Reinforcement Learning (AutoRL): A Survey and Open Problems
• Gaussian Process-Based Refinement of Dispersion Corrections
• Putting Density Functional Theory to the Test in Machine-Learning-Accelerated Materials Discovery
• Scalable Constrained Bayesian Optimization
• Multiple Adaptive Bayesian Linear Regression for Scalable Bayesian Optimization with Warm Start
• Actively Learning what makes a Discrete Sequence Valid
• Parallel Hyperparameter Optimization Of Spiking Neural Network
• Bayesian Graph Neural Networks for Molecular Property Prediction
• Achieving Robustness to Aleatoric Uncertainty with Heteroscedastic Bayesian Optimisation
• Practical Batch Bayesian Optimization for Less Expensive Functions
• GPU Accelerated Exhaustive Search for Optimal Ensemble of Black-Box Optimization Algorithms
• Myopic Bayesian Design of Experiments via Posterior Sampling and Probabilistic Programming
• Domain Knowledge Injection in Bayesian Search for New Materials
• Bayesian Variational Optimization for Combinatorial Spaces
• Bayesian Screening: Multi-test Bayesian Optimization Applied to in silico Material Screening
• Preferential Batch Bayesian Optimization
• Practical Bayesian Optimization of Objectives with Conditioning Variables
• New Heuristics for Parallel and Scalable Bayesian Optimization
• Doubly Bayesian Optimization