Device-independent quantum cryptography

Research Lead: Peter Brown

Both device-independent (DI) and semi-device-independent (SDI) offer very strong security guarantees by removing various levels of trust from the devices on which they operate. Here we investigate the possibilities and limitations of these two approaches to quantum cryptography – ultimately striving to understand their potential as near-term quantum technologies.

Device-independence

In DI cryptography we do not trust the quantum systems used in the protocol. Instead we rely on observing nonlocal correlations (Bell-inequality violations) between two or more non-communicating devices. These nonlocal correlations are necessarily random and hence through their observation we can infer the generation of randomness from the devices. By using nonlocal correlations as certificates of randomness we can then build DI protocols for randomness generation and quantum key distribution (amongst other things).

Unfortunately, the rapid generation of good-quality nonlocal correlations is experimentally very challenging. Moreover, current DI protocols tend to only work well in low noise systems. Nevertheless, the first proof-of-principle experiments for randomness expansion and quantum key distribution have recently been achieved. Inspired by this progress, we aim to demonstrate that DI protocols can be a viable future technology. Through improvements to security proof techniques and methods to compute rates we can reduce the experimental requirements of protocols and push the viability of DI forwards.

Semi-device-independence

Semi-device-independence offers a tradeoff between the strong security guarantees of device-independence and the ease of implementation found in more conventional quantum cryptography protocols. In SDI cryptography we ideally want to add assumptions to our black-box devices that are easy to verify, enable simpler experimental implementations (prepare and measure setups) and make the theoretical analysis tractable. Examples of assumptions proposed in the literature so far include: dimension bounds, energy bounds and bounded distrust.

In our group we will investigate new SDI assumptions, develop new protocols in the SDI framework and collaborate with experimental groups in order to implement them. By applying techniques from the device-independent literature we will also look to improve existing SDI protocols and verify that this is a viable near-term alternative to DI. Finally we will investigate connections between the SDI framework and the problem of verifying the correct functioning of quantum hardware.