PhD Studentship: Self-Supervised Learning for Sparse and Heterogeneous Spatio-Temporal Data

Overview

The School of Computing Sciences at the University of East Anglia offers a full-time PhD studentship focused on self-supervised learning for sparse and heterogeneous spatio-temporal data. The project investigates cutting-edge machine learning methods to identify meaningful patterns from complex spatial and temporal data without relying on large labelled datasets. Research directions include representation learning, graph neural networks, masked learning, transfer learning, and methods for integrating heterogeneous observations. The work initially connects with a Natural Environment Research Council funded project predicting microbial contamination in bathing waters, with broader applications in healthcare, environmental monitoring, and infrastructure.

Who can apply

This opportunity is available to United Kingdom applicants. Candidates must hold a standard minimum entry requirement of a 2:1 degree in Computer Sciences, Data Science, or a related subject, and must be curious, motivated researchers interested in developing new machine learning methods.

What you get

The successful candidate receives funding comprising home tuition fees and an annual stipend for three years. The institutional PhD stipend rate is £21,805 per annum, which is subject to any agreed inflationary uplift for the 2027 to 2028 period.

Requirements

Applicants must participate in a competition for a Faculty of Science funded studentship. The position is based on site at the School of Computing Sciences at the University of East Anglia in the United Kingdom and requires full-time study starting on 1 October 2027.

Key dates

The closing date for applications is 2027-02-24. The program officially starts on 1 October 2027.