PhD Studentship: Learning and Decision-Making Under Uncertainty in Public Health

Overview

The Faculty of Engineering at the University of Bristol offers a PhD studentship on learning and decision-making under uncertainty in public health. Decision-making under uncertainty is a core challenge in AI and data science, especially in dynamic settings where observations arrive in sequence and decisions affect future outcomes. The project will develop new machine learning and statistical methods for adaptive learning, sequential decision-making and control, motivated by public health and epidemiology.

The focus is on reinforcement learning, active learning, Bayesian decision theory and stochastic optimisation for partially observed, evolving systems. Key directions are adaptive data acquisition that maximises information gain under resource constraints, reinforcement learning for sequential intervention and control, and robust methods that cope with incomplete observations, changing environments and distribution shifts. The research combines probabilistic modelling, network-based representations and modern AI to make scalable, interpretable decisions in complex systems, with applications such as adaptive disease surveillance, outbreak monitoring, resource allocation and intervention planning using dynamic mobility and contact networks. The student will be supervised by Dr Mengyan Zhang, whose research centres on sequential decision making and public health.

Requirements

The ideal candidate has foundational knowledge of machine learning and strong self-motivation.

What you get

The funding is a four-year university scholarship starting in the 2027/28 academic year, with competitive worldwide funding. It provides a minimum tax-free stipend at the UKRI rate, which for 2025/26 was a standard stipend of 20,780 pounds plus 8,400 pounds in research training support grant, and full tuition fees are covered.

How to apply

Applicants make an online application to the university and also email the supervisor with a CV, a one- to two-page personal statement, transcripts of qualifying degrees, an optional but preferred research proposal and any supporting materials such as research outputs or code. Because of the volume of enquiries, only shortlisted candidates may receive a response.

Key dates

The closing date is 31 January 2027.