Machine Learning With Homomorphic Encryption For Linear Regression Model
Vedullapalli Adarsh, Lisha Varghese · Zenodo (CERN European Organization for Nuclear Research) · 2022
Abstract—Machine learning jobs typically choose to transfer data and models to a third-party server to take use of cloud computing's compute and storage capability. Third-party servers, on the other hand, may experience data privacy breaches throughout the collection, storage, and use of data, particularly in the domains of finance, medical treatment, and biometrics. In the cipher-text domain, homomorphic encryption technology allows for cipher-text calculation without decryption and can be utilized for machine learning model training and prediction.