Publications

In this paper, we perform a detailed review of new federated learning methodologies for addressing challenges with healthcare data. We also give detailed recommendations to help improve the quality of the methodology development.

In this paper, we propose a novel model architecture for domain generalisation, namely a disentangled autoencoder (Dis-AE).
Dis-AE provides a scalable solution for future research on large collaborative medical datasets suffering from domain shift on many domains. This may be particularly useful in multi-centre studies and trials using repeated measurements.