Dr Joe Fitzsimons, Founder and CEO at Horizon Quantum Computing
When asked about the prospect of quantum computers in Singapore and Southeast Asia (SEA), Horizon Quantum Computing Founder and CEO Dr Joe Fitzsimons believes in the possibility of having sufficient infrastructure to take the technology to the next level.
“At the moment, we are building out a facility to host quantum computers just next door to our office here. We will have a space to host up to three quantum computers. It’s just nearing completion, and we should have the first quantum computer up and running early next year,” he tells e27 in an interview.
“Obviously, we are a long way from Europe or the US. So, shipping things becomes a little bit more complicated. But generally, we have not had any difficulty with infrastructure.”
Horizon Quantum Computing is developing a new generation of programming tools to simplify and expedite the process of developing software for quantum computers.
Before founding the company in 2018, Dr Fitzsimons was an associate professor at the Singapore University of Technology and Design, where he led the Quantum Information and Theory group. He was also a principal investigator at the Centre for Quantum Technologies, contributing to theoretical computer science and physics through his research.
How Transparently.AI uses Artificial Intelligence to detect accounting manipulation, fraud
Earlier in his career, he was a fellow of Merton College, Oxford, and a senior research fellow in the Materials Department at the University of Oxford. During this time, he co-invented universal blind quantum computing, which has since become recognised as an important enabling technology for securing cloud-based quantum computing.
Dr Fitzsimons holds a doctorate from the University of Oxford, where his research focused on quantum computing architectures, and a bachelor of science degree in theoretical physics from University College Dublin.
In this interview, he explains how quantum computing will play a big role in accelerating AI innovation in various industries, particularly in SEA. The following is an edited excerpt of the conversation:
Quantum computing is often described as a technology that will revolutionise various industries. How do you see it complementing or accelerating AI developments, especially in SEA?
The confluence of quantum computing and AI is very interesting, and I think it is likely to have a very large impact. However, it is not the first use of quantum computing. So, we know that you can use quantum computing to accelerate the calculations for quite a wide variety of machine learning models, which is important. Because, if you look at the bottleneck with training larger and larger models, as we go from GPT-4 to 4o and so on, the computational part that goes into training these models is enormous.
It is a limiting factor. If we can make it much easier and more energy-efficient to train these models, we cannot only make better models but also do it with a lot less.
Also Read: Global Microsoft outage demonstrates the need for DePIN computing
Now, getting there is challenging, as quantum computing is still pretty nascent. Quantum computers do not yet outperform conventional computers for any task of interest, but they are getting close to that regime.
When I say it will not be the first application of quantum computing, what I mean is that quantum machine learning, for the algorithms that we know, gives an advantage over existing classical models. They tend to need something called quantum random access memory. That is like a quantum version of the RAM in a conventional computer, which has not been demonstrated yet. So, we have not seen someone produce this kind of memory and put it in a quantum computer.
At the moment, they are just essentially large processors. So, that level of technological development still needs to happen if we want to see some of the benefits of machine learning.
There have been a lot of explorations of nearer-term quantum machine learning algorithms that natively exploit how quantum creators process information. But for those, it is far less certain whether there is going to be an impact or not; they may end up not performing as well as existing classical models. However, for many widely used models, we know you can accelerate them on a quantum computer, but it uses this kind of further term model where we need to have error corrected.
If you look at key sectors in Southeast Asia, such as healthcare, finance and logistics, what are the most promising quantum computing applications that could significantly impact these sectors?
Each of these is pretty prominently represented in quantum computing at the moment.
The most obvious examples are based on the fact that quantum computers can efficiently simulate chemistry. It is easy for quantum computers to simulate chemistry in a way that might be quite difficult for conventional computers, at least for some molecules. This includes not just simulating the molecule itself but also chemical reactions. For example, simulating the folding of proteins can be done very faithfully on a quantum computer.
Also Read: The new era of computing: Single board computers for home automation and AI
It opens up a path to start doing experiments much more in simulation rather than in the wet lab, without the problem of divergence between simulation and the wet lab experiment. This opens up ways to things such as drug discovery, as protein folding is relevant for conformal diseases such as Alzheimer’s.
If you move over to finance or logistics, you see a wide range of applications opening up again, particularly in finance. A very dominant computational problem is the Monte Carlo simulation.
Quantum computers are better at Monte Carlo simulations than conventional computers. You only need to simulate the square root of the number of situations you would need for a conventional computer to get the same amount of precision.
Some banks use a tremendous amount of computing power that can be significantly reduced by quantum computation or, equivalently, getting them down to something that can run much more quickly at much higher frequency. That gives you better insight into the overall risk profile and health system.
If we compare SEA to major tech hubs such as the US or Europe, what are the unique challenges and opportunities for quantum computing in this region?
The industry in the region is a little bit more conservative. I would say the US is probably the place where there is the greatest risk appetite in terms of the willingness to experiment with new technologies at an early stage.
Clearly, with a booming population in SEA, there is tremendous tech expansion in the region, so there is a lot of potential for quantum computing here. Singapore has been an early investor in quantum computing. They have had the Centre for Quantum Technologies for nearly 20 years.
Also Read: Silence Laboratories raises US$4.1M for privacy-preserving collaborative computing
So, there has been high-quality research in the region for quite a long period of time. At the moment, I would say, the industry is slightly slower in taking off in Singapore and SEA than in places such as the UK or the US. However, there are also advantages to starting a quantum computing company.
A company like ours is reliant on being able to attract talent, so it is important to have a skilled talent base. There is definitely academic expertise in the area, people who have trained locally or abroad and come back.
However, there are fewer direct competitors as there are only a few quantum computing companies in Singapore. For a company like us, we are also very, very reliant on trade secrets and IP. This means there is less leakage or less potential for leakage between companies. Whereas, if you’re based in Palo Alto or Berkeley, there are a lot of quantum computing companies, and people are moving between them.
What can the government and academic institutions in Singapore and SEA do to foster innovation in the field, especially in the area of talent?
There are different sides to what is needed in industry.
It is not enough to be an expert in quantum computing to actually turn that into a real technology; we also need exposure to a more engineering[—centric] mindset, which is sometimes at odds with the path people take in computing.
So, I came from theoretical physics. It is a somewhat different path, and the mindset can be different. The approach to tackling problems can be different, but if you are trying to create a new technology … then you also need to focus on good engineering practice and repeatability. These are things that are sometimes under-emphasised in academic efforts.
Also Read: Navigating spatial computing: Augmented, virtual, and mixed reality trends
What becomes important is not just people being exposed to the technology at an early stage. There are increasing courses at the undergraduate level that cover the basics of quantum computing. We have many people who come in with a reasonable level of understanding of what technology is. However, if we want to create a workforce in the area, it is important to expose people to what the industry is like during their training. Otherwise, it can be a bit of a shock to make the leap.
How do you see the trend shifting in Singapore and SEA in the next few years?
What I would say about quantum computing is that it started to move very quickly.
When IBM put its first one on the cloud, it was a five-qubit system in 2015 or 2016. Today, they have a 1,100 qubit system. So, the complexity of the devices is growing very, very quickly. Last month, we also saw the first full demonstration of quantum error correction coming out of Google.
Now, how quickly does that mean things will progress? It is difficult to say, but I would say that, within two to three years, we will be in the low noise regime for quantum computing, where the errors can be corrected to a good extent. So, the error rate will become one in a million instead of one in 100.
When we get there … there is a real chance to see meaningful advantages.
The difference we have between now and then, when the first conventional computers were emerging, is that we have the internet now, which means we can roll out this change. The adoption can be much, much quicker because you do not need to go install a computer at everyone’s office or something like that.
—
Image Credit: Horizon Quantum
The post Horizon Quantum CEO on the Singapore advantage in starting a quantum computing company appeared first on e27.