IN FOCUS

Quantum Computation

by Marc Almagro
Photography by Mohammad Izzadely
13 Feb 2018

Given the nature of Dr. Joseph Fitzsimons' research, it is unlikely that most people will ever interact with it. Quantum computing is developing yet these protocols are meant to run behind the scenes, securing computing services without being intrusive.

Dr. Joseph Fitzsimons is a theoretical physicist working on the development of quantum computation. He is currently an Assistant Professor at Singapore University of Technology and Design, and a Principal Investigator at the Centre for Quantum Technologies. Dr. Fitzsimons is a National Research Foundation Fellow, and in 2016 was named an Innovator Under 35 Asia Pacific honouree. Dr. Fitzsimons holds a BSc in theoretical physics from University College Dublin, and a DPhil from the University of Oxford. Prior to moving to Singapore, Dr. Fitzsimons was a fellow of Merton College, Oxford, and a senior research fellow in Oxford’s Department of Materials. Quantum computing protocols play an important role: By hiding delegated computation, blind quantum computing protocols will allow for R&D groups to make use of quantum computing without having to worry about disclosing proprietary or confidential information to their service provider

Portfolio: How far have you progressed in your research on the use of quantum mechanics enhancing the security of networked computation? 

Dr. Joseph Fitzsimons: I first started working on this topic around 2007, while I was finishing my doctorate in Oxford. My two colleagues, Anne Broadbent and Elham Kashefi, and I had cooked up a trick using quantum effects, which we thought might be useful for cryptography.

We spent almost a year considering different uses before we realized that it gave a very natural solution to the problem of keeping hidden a calculation run on an untrusted computer. The protocol we came up with basically allowed a user to run a calculation of their choice on a remote quantum server without revealing anything about the computation to the owner of the server.

At the time, the usefulness of such a protocol was very speculative. So if you read our first paper, the introduction is imagining a future where there are a few quantum computers, but they are expensive and difficult to operate, and so are confined to a few centres around the globe.

Then, we reasoned, it would be useful for users with limited quantum devices to be able to run calculations on these more powerful quantum computers, while still keeping them private. This was pretty speculative at the time, but if you fast forward to 2018, it is becoming clear that this model of centralized quantum computers is being heavily pushed by industry and will likely be with us for quite some time.

What have come up since the time that you first worked on the subject a decade ago? Are we going to see a prevalent use for quantum computers?

Since our first foray into this field, there has been a lot of progress on developing protocols for secure quantum computing. We realized that it is possible to use our ‘blind’ computing protocol as the basis for further secure computing protocols, and in particular showed that, by embedding traps into their calculation, a user could easily check the result of their calculation. This is particularly useful if you think the operator of the remote computer might try to interfere with your calculation, but is also useful in early quantum processors where imperfections in the device may introduce unintentional errors.

I have been fortunate enough to have colleagues on the experimental side who were willing to try to implement some of these protocols, and so in 2012, colleagues in Vienna were able to demonstrate our blind quantum computing protocol in a small-scale test. Quantum computers are still in their infancy, and so the calculations that we hid were all very simple. However, as a proof of concept it was very effective, and we were able to show that the server could learn almost nothing about the hidden calculation. Shortly there after, in 2013, we were able to work with the same group in Vienna to demonstrate our verification protocol in operation, and to use it to certify the outcome of a simple calculation.

Our recent work has focused on trying to make blind computation and verifiable computation more practical, by eliminating some of the drawbacks of our original protocol. We have had quite a lot of success in reducing the amount of back and forth communication necessary for such protocols, which will be important if they are ever to find widespread adoption.

One drawback of our original protocol was that it could not function over the Internet, but required the user to be able to communicate directly with the server, since security relied on messages sent to the server encoded in quantum states of light. This is an issue that we have put quite a lot of thought into recently, and we are beginning to see progress on blind and verifiable quantum computing protocols which really could be run by a user with an ordinary laptop and internet connection anywhere in the world.

My research has focused on ensuring that we can trust such calculations, and I hope it will enable us to move towards that future with confidence rather than trepidation.

Are you expecting practical applications of your research findings in everyday human activities to happen soon? Can you name some of these potential applications? 

Quantum computers are still at a very early stage, comparable to the very first computers developed in the 1940s. While I fully expect quantum computers to become as ubiquitous as conventional computers are today, that will take many decades to come about.

The first applications of quantum computers will likely not be in consumer–facing applications, but rather in industry–facing areas including optimization, machine learning, and simulating chemistry. The ability of quantum computers to efficiently simulate quantum effects gives them a huge advantage for simulating physics, and so areas that rely heavily on simulating matter at a molecular level, such as computational chemistry, materials science and drug discovery are likely to benefit from quantum computers.

Up until this year, existing quantum processors could be directly simulated on a modest home computer of laptop. Recently a number of research groups and technology companies have announced processors which have increased complexity, and may outperform even the most powerful conventional supercomputers for a very narrow range of problems. This doesn’t quite mean we are at the stage of quantum computers being a useful computing resource, since the problems they excel at are unlikely to be useful. However, it does signify an important step in their evolution, and it may not be long before we see the first practical applications start to emerge.

Several industry players have announced plans to grant access to their quantum processors to users over the Internet. In fact, IBM currently offers free access to a number of quantum processors via the web to both researchers and home users. As the technology begins to mature, I hope that our protocols well play an important role in ensuring the security and reliability of such services.

How will they improve the lives of people?

The nature of my research means that most people are unlikely to ever interact with it. Even when quantum computing matures, these protocols are meant to run behind the scenes, securing computing services without being intrusive. They do, however, fill an important role. By hiding delegated computation, blind quantum computing protocols will allow for R&D groups to make use of quantum computing without having to worry about disclosing proprietary or confidential information to their service provider.

Verification protocols will allow users to take advantage of quantum processing even before devices become 100 per cent reliable by making it easy to check the results of any calculation performed on a quantum device. Ultimately, the results of calculations performed on quantum computers will begin to enter peoples lives in ever more diverse ways, from engineering calculations to simulations of drug interactions. My research has focused on ensuring that we can trust such calculations, and I hope it will enable us to move towards that future with confidence rather than trepidation.

I understand that your quantum computing research also impacts such fields as metrology and quantum information theory. What can you tell us about it? How do you findings help architecture and other fields? Please provide us with detailed examples.

The general idea of taking quantum mechanics into account when we think about information and computation gives rise to a field known as quantum information science, which subsumes quantum computing and other areas such as quantum communications.

This quantum approach to information can be harnessed in a number of different ways. In computing it leads to more efficient algorithms, while in communication it can lead to more efficient protocols, which require less communication to accomplish a given task, and to more secure communications protocols. Nature is best described by quantum mechanics, and so quantum information science provides an important set of tools for analysing our interactions with the world around us. This is particularly prevalent in metrology, the science of measurement, where it has been shown that sensors which can maintain quantum states can make much more precise measurements than conventional sensors. This has a host of applications in imaging, and measuring magnetic fields and optical properties of materials.

Although much of my recent work has focused on secure computing protocols, my research is not restricted to this area, and I have made contributions to a number of other areas. Early in my career my main research focus was on designing hardware architectures for quantum computers, to try to overcome some of the challenges faced in building such devices.

Most of my contributions to this area focused on modular designs, where information was stored in a matter system, such as a trapped ion or a colour defect in diamond, with interaction between these systems implemented using optical measurements. Although such systems are not quite as common as the superconducting systems being pursued by IBM and Google, they offer an alternate path to large scale quantum computers. The quantum computer architecture being pursued in Oxford, as part of the Networked Quantum Information Technologies (NQIT) hub, incorporates ideas I had worked on with colleagues there.

I have also worked occasionally on metrology, which makes use of many of the same tools as quantum computing. One of our most interesting results in this area was to show that quantum sensors could make more precise measurements of magnetic fields than conventional sensors, even in the presence of noise (the random errors introduced intermittently by the outside environment).

Before our work it was known that quantum sensors had an advantage, but only in an ideal noiseless setting, and there were some indications that the presence of noise would eliminate any quantum advantage. We managed to show that this was not the case, and that there were strategies which could mitigate the effect of noise in realistic settings. This was work done at Oxford before moving to Singapore.

More recently, I have been using the mathematical tools of quantum computing in more diverse areas. Along with some colleagues at the Centre for Quantum Technologies here in Singapore, I have been using the tools of quantum information to study quantum correlations. These capture the relationship between measurement results obtained at different positions in space, possibly at different times. This provides a way of studying the causal relationships in a series of events at a very fundamental level.

Rather strangely, we found that at a quantum level the adage that correlation does not imply causation is no longer true and that it is possible to find correlations that can only possibly have come from a causal relationship. While this is unlikely to have any practical consequences in terms of technology, it provides a useful theoretical tool for better understanding the fundamental laws of nature.