Jack Buckingham

Summary

I am a PhD student studying Bayesian optimisation, supervised by Professor Juergen Branke (WBS) and Professor Ivo Couckuyt (Ghent University; imec). I am currently working on a form of robust optimisation called ‘active robustness’.

While my focus is on designing new Bayesian optimisation algorithms, I am particularly interested in problems motivated by “green” applications. That is, things which solve some of the problems created by climate change and more generally by human impact on the environment.

My Research

Bayesian optimisation is a technique for maximising an expensive, unknown function using only a small number of evaluations. For example, consider optimising the parameters of a simulation which takes minutes/hours to run or tuning the hyper-parameters of a neural network which must be retrained to test each new set.

An orthogonal concern in engineering design is that of robustness. Designs often need to perform well in a variety of environmental conditions and must not be over-specialised to any one target environment. Active robustness refers to the bi-level optimisation problem of designing a robust product which has some capacity to adapt to its environment in real time. My research combines these two themes.

Publications & Preprints

Buckingham, J. M., Couckuyt, I., & Branke, J. (2024, Preprint). Bayesian Optimization for Non-Convex Two-Stage Stochastic Optimization Problems. arXiv. https://arxiv.org/abs/2408.17387

Buckingham, J. M., Rojas Gonzalez, S., & Branke, J. (2024, Preprint). Knowledge Gradient for Multi-Objective Bayesian Optimization with Decoupled Evaluations. arXiv. https://arxiv.org/abs/2302.01310v2

Hart, W. S., Buckingham, J. M., Keita, M., Ahuka-Mundeke, S., Maini, P. K., Polonsky, J. A., & Thompson, R. N. (2024). Optimising the timing of an end-of-outbreak declaration: Ebola virus disease in the Democratic Republic of the Congo. Science Advances. 10, eado7576 DOI: 10.1126/sciadv.ado7576

Selected talks and conferences

23rd September 2024 "Bayesian Optimisation for Non-Convex, Two-stage Stochastic Optimisation Problems" (poster), Heidelberg Laureate Forum. https://youtu.be/1-aTRGvJzPQ?feature=shared&t=2567
10th July 2024 "Bayesian Optimisation for Non-Convex, Two-stage Stochastic Optimisation Problems", University of Warwick, AMP Conference.
15th March 2024 "Making use of decoupled objectives in Bayesian optimisation", Heriot-Watt University, MINDS student seminar.
7th December 2023 "Exploration vs Exploitation: The art of acquisition functions in Bayesian optimisation" University of Warwick, SPAAM seminar. https://youtu.be/EnRzplZ7aDg?feature=shared
19th April 2023 "Bayesian optimisation for active robustness" and "High-dimensional Bayesian optimisation", University of Warwick, Bayesian optimisation workshop with GE.

Education and Experience

2022-present: PhD in Bayesian optimisation, University of Warwick
Working title: Designing actively robust products with Bayesian optimisation

SPAAM seminar organiser 2022/23
SIAM-IMA chapter vice president 2022/23
SIAM-IMA chapter president 2023/24
AMP24 conference committee 2024
SIAM-IMA chapter events officer 2024/25

2023 (June-August): Technical Mentor, Data Science for Social Good UK
Technical mentor to two teams of four fellows working on three-month data science projects
Oversaw the teams' data science strategies and provided guidance on software best practices.
Website: DSSGx2023

2021-2022: MSc in the Mathematics of Real World Systems, University of Warwick (Distinction)
My MSc project was in multi-objective Bayesian optimisation.

2017-2021: Data Engineer/Data Scientist, Pace Revenue
Early member at a dynamic pricing start-up in the hospitality industry.
I worked as both a data engineer and data scientist. Selected projects include using SARIMA models to forecast hotel bookings, and incorporating booking forecasts in the pricing algorithm.

2015-2017: Metrology Engineer, Renishaw
Precision measurement company for the manufacturing industry.
Prototyping new calibration methods; Quantification and propagation of uncertainty; Non-linear optimisation; Bayesian modelling.

2011-2015: Master of Mathematics (BA/MMath), University of Cambridge (Merit)

Headshot

Contact

jack dot buckingham at warwick dot ac dot uk

Office: D2.11 (zeeman)

SIAM-IMA

This year I am an events officer for the Warwick SIAM-IMA chapter. We run the SPAAM seminar series, as well as organising social events, hackathons and the AMP conference.

If you are a Warwick student with an idea, then send us an email and we'll make it happen! If you are from a SIAM chapter at another university looking to collaborate then you can contact us at siam at warwick dot ac dot uk

We're very friendly!

DSSG

Data science for social good (DSSGxUKLink opens in a new window) is an excellent opportunity for developing data science skills on real-world projects with a positive social impact.

Media UN-REDD blogLink opens in a new window, IFORS articleLink opens in a new window

Useful Links

Climate change

Social good