Speaker: Zhaofei Tian

Abstract:
Lattice reduction aided methods have been successfully used in many cryptographic applications. In this presentation, we present an enhanced Jacobi (short as EJacobi) method based on the Jacobi method introduced by S. Qiao in 2012. We compare the EJacobi method with the widely used LLL algorithm. Our experimental results show that measured by orthogonality defect and condition number, the EJacobi method performs better than the LLL algorithm. In terms of running time, the EJacobi method is slower than the LLL algorithm when the dimension of the lattice basis is larger than 30.

Speaker: Dragan Rakas

Abstract:
Fully homomorphic encryption allows computations to be made on encrypted data without decryption, while preserving data integrity. This feature is desirable in a variety of applications such as banking, search engine and database querying, and some cloud computing services. In this presentation, we discuss a method of encrypting circuits and executing encrypted assembly instructions, by combining fully homomorphic encryption and digital logic theory. We use the MIPS Assembly Architecture as a base for our work, and the result is essentially an encrypted programming language, where a remote server is capable of executing program code that was written and encrypted by a local client.