We are currently looking for a motivated PhD researcher to join our team and contribute to an interdisciplinary project focused on understanding the extent to which network-based mean-field models can be used to infer network- and disease-related parameters. The project will study and address several challenges related to (i) structural and practical identifiability (ii) data compression and model-data mismatch (e.g. mean-field models operate at population-level but data often comes at individual-level), (iii) feasibility and performance of different inference scheme, and (iv) portability and scalability of methods to models that encode network properties beyond basic metrics.

From a methodological viewpoint, the PhD project will consider both controlled experiments where network and data quality and resolution are controlled and well-known, and real-world scenarios. The former will be the testbed to explore the usefulness and limitations of inference schemes such as the classical maximum likelihood approach or the recently proposed dynamical survival analysis method using mean-field models of increasing complexity with higher fidelity in terms of network encoding. There is further scope to develop custom inference schemes from a rigorous statistical viewpoint or to improve the performance or scalability of optimisers used in parameter searches.