Research

Cryptography is defined as the study of techniques to secure data, whether it is stored or in transit. Research in cryptography helps us understand and design tools to address security in a variety of real-world applications, many of which we use every day, as well as thinking of future security.

Currently, the Cryptography Group at Royal Holloway is particularly interested in developing solutions in four important areas: secure computation, post-quantum cryptography, protocol security, and cryptography for AI.

Secure Computation

Currently deployed cryptography mainly deals with protecting confidential data at rest and during transit. But when data needs to be processed, then it must be decrypted, making it vulnerable to attacks.

We therefore look at how to overcome this issue by advancing research in secure multi-party computation and fully homomorphic encryption. Both concepts provide a provably secure way to process data under encryption and therefore are a key enabler for collaborative and outsourced computation on highly sensitive data.

The research of our group in this area is primarily focused on improving the efficiency of encrypted computing to enable building privacy-preserving applications at scale, for example, for secure distributed database analytics, which is required in a number of use cases such as enforcing anti-money laundering policies.

Selected Publications

Post-Quantum Cryptography

Currently deployed public-key cryptography relies on certain hard problems, such as that it is hard to factor a large product of primes into its two prime factors. These problems are hard to solve with existing computers, but if a large-scale quantum computer is built, they would be significantly easier, to the point that existing deployments would be considered insecure.

In the Cryptography Group, we are working on developing and cryptanalysing so-called post-quantum cryptographic schemes that can replace existing public-key cryptography. Post-quantum schemes are based on different problems, that are believed to be hard to solve using both classical and quantum computers. This ensures we can deploy a replacement to existing vulnerable schemes.

Selected Publications

Protocol Design & Analysis

Cryptography is the cornerstone of security online and is becoming exceedingly important as our lives become increasingly digital. At the heart of this are a myriad of cryptographic protocols, all with their own unique forms. While often in the background, these have gained more public awareness, mainly driven by legislation proposed by various governments.

We endeavour to understand the security of the protocols by providing a mathematical foundation for their security. We do not only examine currently deployed schemes, but we are also looking forward and examining schemes that are slated for future deployment.

These schemes cover a wide spectrum of areas ranging from day-to-day, which sometimes get overlooked, such as protocols that protect your data while browsing the web to secure messaging and end-to-end encryption, as well as more complex protocols such as secure electronic voting schemes and emergent technologies such as blockchains and smart contracts.

Selected Publications

Cryptography for AI

Emerging AI services are set to revolutionize the way we work and live. However, the vast amounts of data involved in AI training and inference bring substantial privacy and security challenges, including risks of adversarial interference and the need for intellectual property protection.

The Cryptography Group tackles these issues by developing efficient, secure computation methods for AI training and inference. We also explore techniques like watermarking to ensure the protection of intellectual property in AI systems.

Selected Publications