Home > Research > Publications & Outputs > General-purpose Information-theoretical Bayesia...

Electronic data

  • 2021mossphd

    Final published version, 6.76 MB, PDF document

Text available via DOI:

View graph of relations

General-purpose Information-theoretical Bayesian Optimisation: A thesis by acronyms

Research output: ThesisDoctoral Thesis

Published
Publication date2021
Number of pages213
QualificationPhD
Awarding Institution
Supervisors/Advisors
Award date15/04/2021
Publisher
  • Lancaster University
<mark>Original language</mark>English

Abstract

Bayesian optimisation (BO) is an increasingly popular strategy for optimising functions with substantial query costs. By sequentially focusing evaluation resources into promising areas of the search space, BO is able to find reasonable solutions within heavily restricted evaluation budgets. Consequently, BO has become the de-facto approach for fine-tuning the hyper-parameters of machine learning models and has had numerous successful applications in industry and across the experimental sciences.This thesis seeks to increase the scope of information-theoretic BO, a popular class of search strategies that regularly achieves state-of-the-art optimisation. Unfortunately,current information-theoretic BO routines require sophisticated approximation schemes that incur substantially large computational overheads and are, therefore, applicable only to optimisation problems defined over low-dimensional and Euclidean search spaces. This thesis proposes information-theoretic approximations that extend theMax-value Entropy Search of Wang and Jegelka (2017) to a much wider class of optimisation tasks, including noisy, batch and multi-fidelity optimisation across both Euclidean and highly-structured discrete spaces. To comprehensively test our proposed search strategies, we construct novel frameworks for performing BO over the highly-structured string spaces that arise in synthetic gene design and molecular search problems, as well as for objective functions with controllable observation noise. Finally,we demonstrate the real-world applicability of BO as part of a sophisticated machine learning pipeline for fine-tuning multi-speaker text-to-speech models .