Abstract: This paper proposes the application of logic locking over TFHE to protect both user data and algorithms, such as input user data and models in machine learning inference applications. With ...