Background

Privacy-Preserving Machine Learning (PPML) involves developing methods and technologies that allow machine learning models to be trained and applied on sensitive data while ensuring that the privacy of the data is maintained. This field combines advancements in cryptography, data privacy, and machine learning to address the growing need for secure data handling in various applications.

Research Goals

This project will propose new cryptographic algorithms and methods that ensure the privacy of sensitive data while allowing for effective machine learning model training and inference. It will integrate techniques such as homomorphic encryption, secure multi-party computation (MPC), federated learning, and differential privacy into machine learning frameworks.

Skills Needed

Familiarity with C++, Python, Cryptography, Machine learning.

Skills Gained

Learning how to incorporate privacy-preserving methods into machine learning models and workflows, enhancing skills in training machine learning models, especially under privacy constraints, and evaluating their performance.