← Toutes les offres
UO

PhD Studentship: Machine Learning and Uncertainty Quantification for Complex Scientific Systems

University of Birmingham

Birmingham
CDD
Temps plein
Biological Sciences
Publié le 20 novembre 2026

Description du poste

Machine learning is transforming how we study complex systems across science, engineering, and medicine. However, many real-world processes are inherently uncertain, dynamic, and only partially observable. Scientific measurements are often noisy, incomplete, or collected at different spatial and temporal scales. Developing reliable mathematical and computational methods to learn from such data is an important challenge at the intersection of applied mathematics, statistics, and artificial intelligence. We are seeking motivated PhD students interested in developing new mathematical, statistical, and machine learning approaches for modelling, understanding, and predicting complex scientific systems. The research will focus broadly on scientific machine learning, uncertainty quantification, stochastic modelling, and data-driven inference. Potential directions include learning from noisy and high-dimensional observations, modelling complex stochastic dynamics, quantifying uncertainty in data-driven predictions, and developing reliable statistical methods for scientific applications. A key motivation is to bridge the gap between modern machine learning and the mathematical principles needed for trustworthy scientific inference. Rather than treating machine learning as a black-box prediction tool, the project aims to develop methods that can capture uncertainty, reveal underlying dynamics, and provide meaningful insights into complex processes. Potential applications span several interdisciplinary areas, including: Biology and medicine: Understanding cellular and molecular dynamics, analysing noisy biological measurements, and studying complex biological processes. Physics and engineering: Learning stochastic dynamical systems, reconstructing physical processes from imperfect observations, and improving scientific modelling. Scientific data analysis: Developing robust methods for high-dimensional, time-dependent, and spatially structured data. The precise research direction will be shaped by the successful candidate's interests and background. There will be opportunities to explore both fundamental methodological questions and applications involving real scientific data, with an emphasis on developing mathematically principled approaches that address meaningful practical challenges. We welcome applicants with a strong background in applied mathematics, mathematics, statistics, machine learning, computational science, or closely related disciplines. An interest in probability, stochastic processes, statistical inference, optimisation, or scientific computing would be particularly valuable. Applicants should be enthusiastic about mathematical and computational research and interested in developing new methods motivated by real-world scientific problems. Prior experience in biological or medical research is not required. The project offers an opportunity to work at the interface of mathematical theory, modern machine learning, and interdisciplinary scientific applications, contributing to the development of reliable tools for understanding uncertainty in complex systems. The project will be supervised by Dr Mingtao Xia. Funding notes: Tuition funding will be available for UK-based applicants, but outstanding international students may also be considered.