Making better use of health data without sacrificing privacy

UvA is coordinating a new seven million euro European project which is going to develop privacy preservation techniques for health-related data. UvA-researcher Francesco Regazzoni from the Complex Cyber Infrastructure groep of the Informatics Institute received a grant of € 750.000 to lead as project coordinator the new seven million euro European project SECURED This project aims to scale up techniques for anonymization, collaborative computation and synthetic generation of health data so that they can safely be used in medical applications. Currently, such techniques are too impractical for a number of applications, such as real-time image classification, genomic analysis, or remote monitoring of patients. They are also often too unreliable to be used as training materials in education or for machine-learning training. Synthetic data generation is one of the technologies explored in the project. 'If you want to study a disease that is extremely rare and use the support of machine learning for such study', says Regazzoni, 'you will need a large amount of training data, but they are not always available. That's an example of a case in which you want to generate realistic data synthetically.