Preferential Batch Bayesian Optimization

Eero Siivola, Akash Kumar Dhaka, Michael Riis Andersen, Javier González, Pablo García Moreno, Aki Vehtari

Research output: Chapter in Book/Report/Conference proceedingConference contributionScientificpeer-review

Abstract

Most research in Bayesian optimization (BO) has focused on direct feedback scenarios, where one has access to exact values of some expensive-to-evaluate objective. This direction has been mainly driven by the use of BO in machine learning hyperparameter configuration problems. However, in domains such as modelling human preferences, A/B tests, or recommender systems, there is a need for methods that can replace direct feedback with preferential feedback, obtained via rankings or pairwise comparisons. In this work, we present preferential batch Bayesian optimization (PBBO), a new framework that allows finding the optimum of a latent function of interest, given any type of parallel preferential feedback for a group of two or more points. We do so by using a Gaussian process model with a likelihood specially designed to enable parallel and efficient data collection mechanisms, which are key in modern machine learning. We show how the acquisitions developed under this framework generalize and augment previous approaches in Bayesian optimization, expanding the use of these techniques to a wider range of domains. An extensive simulation study shows the benefits of this approach, both with simulated functions and four real data sets.
Original languageEnglish
Title of host publicationProceedings of 31st International Workshop on Machine Learning for Signal Processing, MLSP 2021
PublisherIEEE
Pages1-6
Number of pages6
ISBN (Electronic)978-1-7281-6338-3
ISBN (Print)978-1-6654-1184-4
DOIs
Publication statusPublished - 15 Nov 2021
MoE publication typeA4 Article in a conference publication
EventIEEE International Workshop on Machine Learning for Signal Processing - Gold Coast, Australia
Duration: 25 Oct 202128 Oct 2021
Conference number: 31
https://2021.ieeemlsp.org/

Workshop

WorkshopIEEE International Workshop on Machine Learning for Signal Processing
Abbreviated titleMLSP
Country/TerritoryAustralia
CityGold Coast
Period25/10/202128/10/2021
Internet address

Keywords

  • Uncertainty
  • Machine learning
  • Signal processing
  • Data models
  • Bayes methods
  • Noise measurement
  • Usability

Fingerprint

Dive into the research topics of 'Preferential Batch Bayesian Optimization'. Together they form a unique fingerprint.

Cite this