PREMAL: Practical Encrypted Machine Learning

Machine learning has the potential to unlock new applications across many sectors, but its use on sensitive data is often constrained by privacy regulations and data-sharing concerns. Privacy-Preserving Machine Learning (PPML) aims to address these challenges.

PREMAL focuses on fully homomorphic encryption (FHE), which enables computations to be performed directly on encrypted data, ensuring that sensitive information is never exposed in plaintext. While promising, the practical use of FHE in machine learning remains limited.

The project aims to develop a comprehensive methodology for encrypted machine learning, addressing key gaps such as encrypted inference, training on encrypted data, and preventing data leakage in federated learning. The outcome will be machine learning systems with strong, built-in privacy protections suitable for real-world deployment.

Funding

This project is funded through the Research council of Norway's funding scheme “Researcher projects for ICT Renewal and Development” (forskningsradet.no).

Partners

• Simula UiB (Norway), coordinator
• SimulaMet (Norway)
• FFI (Norway)