Efficient and Principled Scientific Discovery through Bayesian Optimization: A Tutorial

TL;DR

This tutorial introduces Bayesian Optimization as a systematic, probabilistic framework to enhance scientific discovery by automating experiment design and resource allocation across various scientific domains.

Contribution

It presents a comprehensive overview of Bayesian Optimization, its core components, workflows, technical extensions, and real-world applications in scientific research.

Findings

BO effectively automates experiment selection in scientific discovery.

Case studies demonstrate improved efficiency in catalysis, materials science, and molecule discovery.

Extensions like batched experiments and human-in-the-loop enhance BO's practical utility.

Abstract

Traditional scientific discovery relies on an iterative hypothesise-experiment-refine cycle that has driven progress for centuries, but its intuitive, ad-hoc implementation often wastes resources, yields inefficient designs, and misses critical insights. This tutorial presents Bayesian Optimisation (BO), a principled probability-driven framework that formalises and automates this core scientific cycle. BO uses surrogate models (e.g., Gaussian processes) to model empirical observations as evolving hypotheses, and acquisition functions to guide experiment selection, balancing exploitation of known knowledge and exploration of uncharted domains to eliminate guesswork and manual trial-and-error. We first frame scientific discovery as an optimisation problem, then unpack BO's core components, end-to-end workflows, and real-world efficacy via case studies in catalysis, materials science, organic synthesis, and molecule discovery. We also cover critical technical extensions for scientific applications, including batched experimentation, heteroscedasticity, contextual optimisation, and human-in-the-loop integration. Tailored for a broad audience, this tutorial bridges AI advances in BO with practical natural science applications, offering tiered content to empower cross-disciplinary researchers to design more efficient experiments and accelerate principled scientific discovery.