PhD in Electrical Engineering

Homomorphic Encryption (HE) is an exciting, advanced type of encryption, which allows computations on encrypted data, without use of a decryption key. Introduced in 2009, such a powerful encryption tool enables privacy-preserving data analysis, however HE suffers in terms of performance due to high computational complexity and particularly memory management demands. Hardware acceleration and optimisation of homomorphic encryption has demonstrated successful speed up factors of over 100x for homomorphic encryption. Moreover, such HE schemes are often based on lattice based cryptography, relying on security via added noisy error vectors, which allow for potential approximation. Further research is needed to investigate the hardware acceleration of homomorphic encryption, balancing performance, security, approximation, and accuracy, to facilitate high performance implementations at the edge.