The future of quantum
What does the future of quantum look like? Vedran Dunjko (Assistant professor at LIACS and LION) is co-founder of the applied quantum algorithms (aQa) leiden interdepartmental initiative research that will investigate this. They are part of the Quantum Delta NL consortium that was recently awarded substantial funds from the government via a so-called 'growth fund'.
How will your research influence my personal life in the future?
‘Behind everything you use every day that has 'smart' in front of its name is a machine learning gadget. Things work better now than they did 15 years ago because there is machine learning behind them. And in 15 years' time, your smart devices will be even better, because they have quantum technology working in the background. Think of even smarter phones that do much of the work for you, and smart cars which make accidents things of the past.'
'Eventually, I hope, there will be a smart world, where everything is integrated and optimised. And for scientists, I hope we get machine-assisted research and AI-assisted science to achieve breakthroughs that are impossible now.'
What exactly does your research involve?
'We design quantum algorithms that can run on quantum computers available in the near future. We are working together with companies such as Shell, Google, Volkswagen and Total. We want to improve the performance of AI by using quantum computers, but also to use of machine learning and AI techniques to improve near term quantum computers. In the end this will enable us to better understand complex data.’
What is the advantage of using quantum technology in AI?
‘Quantum computers can help us perform certain previously unthinkable calculations efficiently. We are now trying to build a code that can be used for handwriting recognition or cancer detection, for example. And we are investigating how different quantum-computable functions are from neural networks. A neural network is made up of artificial units that use a computer programme to imitate the characteristics of real biological neurons. We are now examining whether quantum machine learning models are actually better than the deep neural networks that we currently use in some cases. Ultimately, the crucial question is whether we can find an algorithm that we can actually run on a quantum computer and that makes sense in the real world. But for that we first need a really good quantum compute, and we are slowly getting there.'
Why is the government investing millions in this research?
‘The government's money goes to various pillars of quantum technology. For example, fellow researchers will continue to build the quantum network.’
‘Quantum computing has so far been studied from a very theoretical standpoint. Two years ago, Google made an important breakthrough with the so-called 'quantum supremacy experiment'. They created a quantum device that can do well-defined, programmable calculations that no computer can do in a reasonable amount of time. This is a breakthrough because it can do something in a millisecond that would take an existing computer 10,000 years to do. The problem with this device is that everything has to be customised. In order to build a quantum computer in the next 5 to 10 years, we need to bring together the people who build computers and the people who design algorithms.’
‘And you have to know exactly what you want to do with the algorithms. For that we need industrial partners, people who tell us: these are our real problems and this is what we want to do. The government investment is intended to bring these three players together: the builders, the programmers and the end users to make useful quantum algorithms. You need all three of these groups, preferably under one roof.'
'The strength of Leiden is that we have brought physics, computer science, chemistry and mathematics together in aQa. LIACS is an applied computer science institute that is open to quantum technology. This is quite unique, because not many institutes think about quantum computers as a technology that can solve practical problems in the real world.'
What will LIACS do with the money from the growth fund?
'The part of the money that comes to LIACS is spent on PhD students and post-docs. In Leiden we do not build quantum computers, but train people who can get the quantum computers to do useful things, but also learn to talk to industrial partners and experimentalists. The idea behind Quantum Delta NL is to link students to industrial partners and to the researchers in those companies. The student has to be really committed to really understand what is happening in the company and what is needed. And we hope the companies will also commit and help fund the PhDs.'
What do you hope to have achieved in the next 5 to 10 years?
'That there is a first example in the real world of a broadly beneficial quantum computation. That people in the real world tell me 'this has really improved our lives'. I think the first examples will happen in about 6 to 7 years, although it remains difficult to predict. It depends mainly on the development of the devices and other software breakthroughs.'
‘Quantum computers offer immense perspectives and we must now have ideas about where these currrent quantum computers should develop to next first. What numbers do we need to move forward? That is perhaps the most important result I hope to achieve in three years' time. People in quantum computing really need to sit down and calculate numbers, and then when they say we need 150 qubits and 50,000 logic gates to solve a particular problem, suddenly it becomes real. It may be an immense goal, but when you can attach an economic benefit to it, for example, it becomes very real.'
Interview by: Marcel Tichelaar