The application for the project titled "Privacy compliant health data as a service for AI development (PHASE IV AI)", within the scope of the HORIZON-HLTH-2022-IND-13-02 call, by Sabancı University Faculty of Engineering and Natural Sciences Dean Erkay Savaş and Faculty Member Kamer Kaya has been accepted.
The total budget of the project, which includes 20 partners, including Sabancı University, is 7,599,713 Euros. In the project, the Sabancı University team will provide support in the areas of cryptography, secure multiparty computation (SMPC), homomorphic encryption, federated learning, and privacy-protected artificial intelligence.
The aim of this project is to further advance the most developed existing data synthesis methods in order to create a more generalized approach to the generation of artificial data to be used in the development of artificial intelligence solutions. Because the data is private and distributed, new protocols will be developed with multiparty computation methods based on real data, but without directly using real data, and new metrics will be created for testing and verification.
The project aims to develop privacy-protected synthetic data synthesis methods for different types of data, including electronic health records (EHR) and medical images. In addition, large decentralized data pools will be created that developers of artificial intelligence solutions can use, easily access, and shape according to their needs, again with privacy protection.
With the implementation of the project, a data center will be established in the European Health Data Space, the data market and data service system will be integrated, and the project outputs will be tested and validated on real-world problems, focusing especially on high-impact diseases such as cancer.