2026 · Featured
Forecasting Malaria Incidence in Rwanda with Neural CDEs
Doctoral research: continuous-time deep learning for irregularly sampled malaria data, integrated with optimal control tools for intervention planning.
Overview
My PhD project at the University of Łódź (2025–2029), supervised by Prof. dr hab. Andrzej Nowakowski with Dr hab. D. Machowska as secondary supervisor. The goal is a forecaster that handles the irregular sampling of real epidemiological data and can be coupled to optimal control to recommend intervention schedules for national malaria programmes. A Neural Controlled Differential Equation forecaster is currently under development on top of an established benchmark.
Highlights
- Rolling-origin forecasting benchmark (naive, ETS, AR, pooled ridge) as the baseline for the Neural CDE forecaster
- Data access granted by the DHS Program (Rwanda DHS 2014–15, 2019–20; MIS 2013, 2017, 2023)
- Surveillance data request in progress with the Rwanda Biomedical Centre
- Presented at the XXXI National Conference on Applications of Mathematics in Biology and Medicine, Kościelisko, Sept 2026