Privacy Preserving Machine Learning (PPML)

The project is dedicated to demonstration of Privacy Preserving Machine Learning (PPML) in EO data analytics as an emerging field in data science that strives to address protecting data privacy and confidentiality in a variety of machine learning applications. It is aimed to research and demonstrate secure frameworks for AI/ML, including cryptographic techniques such as Federated Learning (FL), Secure Multi-Party Computation (MPC), Homomorphic Encryption (HE), and Differential Privacy (DP), as well as hardware-based approaches (TEE – Trusted Execution Environment) to provide computing environment suitable for handling and protecting confidential data as well as AI models from disclosure or interference.