Transcript
Rene – Hi! Welcome to QuBites, your bite-sized pieces of quantum computing. My name is Rene from Valorem Reply and today we're going to talk about Sovereign Cloud Quantum Computing and for this I'm honored to have a special expert guest today Florian Neukart. Hi Florian and welcome to the show. How are you today?
Florian – Great, thanks for having me.
Rene – Awesome! Well, first of all Florian, can you tell us a little bit about your background and you know what you're working on and that kind of stuff?
Florian - Yeah absolutely. So, I'm starting with my academic background that's computer science and physics but I've been in the industry for well it almost decades now so for almost 20 years and in industry, I've always been concerned with making the bridge or building the bridge between fundamental research and application so how do I get research results into products and how can we make products that people are excited about, that they can use, that are scalable, that are secure, everything that comes along with whatever a good product has. One of the things I'm very excited about, of course, is quantum Computing and so that's also my academic background and that's where I still work at the University of Leiden in Quantum Computing and at Terra Quantum where I oversee all product development their Quantum is a Quantum technology company and amongst others we work on Quantum computing.
Rene – Awesome, well this is great and like you're saying it's I think really fantastic and very unique, you know, having this academic background and bridging it and bringing this with the business together it's an important part because you know if it just lives in the lab like we need it as products so we got to bridge it and build a bridge kind of as well. So let's dive into today's topic. I saw that you're also working on a Quantum as a service solution or qaas as we say for the German Ministry of Economic Affairs and Climate Action and you're going to construct a platform for Quantum Computing applications for the German industry. So can you tell us what it is and why there's actually need for this for the industry there?
Florian – Yeah, absolutely. So, let's start maybe with Quantum Computing. So as you know Quantum Computing is a not only a very interesting technology but it will change many of the things that we do today or influence at least many of the things that we do today and so it will also modernize or contribute to modernizing most Industries. So be that medicine, so how we develop drugs, so pharmacology so that's one one of the areas that will strongly benefit from it, all material science, so how, for example, we develop new batteries or simulate battery chemistry that's something we can do with quantum computers, all machine learning, all artificial intelligence and so at the moment we're still trying to find out how we can better train artificial intelligence models and how they can or how we can enable them to learn better from data and that's also where quantum computers will strongly be supported as you know machine learning or artificial intelligence in a broader sense is used throughout the industry. So there is no limitation in terms of scope and optimization. It's also one of the fields that will strongly benefit from Quantum Computing. So that examples are traffic flow optimization, Logistics problems Parts distribution, and production and there are many more and so this is why we want to encourage the use of this new technology. But in order to use it, it must be accessible and available and that's what the aim of this project is. So the German Quantum Cloud will provide a hybrid Quantum Cloud so meaning the best of high performance Computing that we know plus Quantum Computing to Industries and Industry Partners can then experiment and try out use cases, bring in their use cases. It's also about exploring where this technology can be best applied. So very often as an industry player you may think Quantum Computing that's way too advanced. For me, it's not something I want to use in my everyday life but it is not correct in our point of view. So there are so many areas and so many use cases that will strongly benefit from Quantum Computing, but again, that's why we want to offer this or bring this technology through the German Quantum Club to Industry and allow them to not only play around but really think about where can I use it, how can I best deploy my applications.
Rene - That sounds great and like you're saying access to Quantum technology is quite limited in the sense that if you want to use a true quantum computer although they are still in this infancy, they're super expensive, right, so and like yeah if you wanted to do a research project in your company you might just not want to invest all of this. But if you offer this as a Quantum as a service, surely that makes a lot of sense.
Florian – Yeah, absolutely agree. So, it really depends on what technology you look at, but all in all none of quantum computing if you were to buy it on-premises is cheap. So that's why availability through cloud systems absolutely makes sense and what's more, Quantum Computing so specifically when we look at native quantum processors, they are very fast. So depending on what problem you solve, it's not that you would take hours or even days to run a problem and get a solution. They're really fast, so that's why if you provide it through a cloud even if you have a limited number of systems available only you can serve a multitude of customers and solve a multitude of problems. So, it's not really that everyone needs a dedicated system in their data center locally, cloud access therefore definitely makes sense also when you think about combining it with classical high-performance computing which lives in the cloud these days. So it also makes sense to bring Quantum Computing to the cloud and the combination of these two worlds, so not only making it available in a cloud so that we have both available but really combining it and integrating these two, that is where the future is, it's not only a bridge technology when we look at today's Quantum chips they are small they are error-prone so it's really systems that require further development, further research sometimes but if you combine it with high-performance computing that's not the case you can solve problems at full scale and that's what industry needs. So the industry of course is always keen on experimenting too but in the end what counts is can you bring it into production, what's the cost of integration and does it save me any money or does it make me more efficient if we can do that which we aim to do with the German Quantum Cloud.
Rene – Makes, a lot of sense, and fully agreed. Let's come back to the title of the episode which is Sovereign Cloud Quantum Computing and now I got it right and you have been you have been living in the US for many years but you're also very well connected to European not just the quantum Computing market but, of course, quantum computing as well and so I think you have a good overview actually about what's going on both continents if you will, what is the main difference you're seeing between Europe and the United States when it comes to the quantum computing business funding and also why is Sovereign Computing important in that context?
Florian – Yeah, so maybe I'll start with funding. So one of the differences that I see is I mean that's not limited to Quantum Computing, it's to high technology investments in general. So one of the differences I see is the amount of money that's spent by VCs, by even individuals high net worth individuals on a specific technology that really want to convert technology to fruition. So that's different I feel in the US but Europe is catching up. So it's not that we can really strongly distinguish between us as all the funding and Europe has none of it, Europe is catching up there's more and more funding going into Quantum Computing, going into other important technologies but that's still one of the distinguishing factors. I think there's easier access to money in the United States and that comes partly from the willingness to risk money. So you as a European, I'm from Austria so I think sometimes we tend to be more risk-averse and really want to be hundred percent sure that an investment will pay off which is something we look at the Silicon Valley in general it's not something that is true here people really risk their money knowing that things could go wrong and what's more it's not the end of the world if things go wrong. Yes of course as an investor you never want to lose money but if you lose money then sometimes that happens but you still go on and try the next thing and even in the same area, the same technology. So it's not the end of the world. So that's one of the big differences I see, I think, in terms of sovereignty every country no matter if it's Quantum Computing or any other advanced technology needs to have access to that technology because otherwise it will create an imbalance in the world and that's not something that any of the companies and any of the governments really would want. So therefore I think Investments need to happen locally as well. Our company for example here at Quantum and subsidiary Qmware, they're European headquarters, they're purely European funded so they are, it will always be anchored in Europe and are less European Quantum technology companies. So I really think Investments like that and also into other technologies such as AIum, they pay off for the local industry and they pay off for the local governments and in the end as I said I think in my point of view it's also about a distribution of power. So to say because otherwise we would see a strong imbalance in the world.
Rene - Yeah, I fully agree and yeah I also see the same thing as like that Europe is actually catching quite up when it comes to Quantum Investments. I mean we the first question right it's just one signal as well right that Germany for example is one country that is investing for the industry into a Quantum as a service solution but also the European commission does a bunch of funding and also a lot of interesting research facilities of course, right, all over Europe and yeah I think we need to use the strengths of, you know, in the European mindset, if you will, to also you know come together and work on a lot of these problems.
Florian - Absolutely and also if I may add to that. I mean the research that happens in Europe, it's absolutely fantastic. It's not only competitive, I would say sometimes best in the world and specifically when you look at Quantum Computing or generally quantum physics and Quantum technologies. So look at the recent Nobel Prize that was of course shared between U.S France and Austria but so Europe has great research going on and I think sometimes the only thing or we're a little behind is the transition from research into industry into products into the commercial level market.
Rene - Yep agreed. So that this always needs to be accelerated a little bit. All right, let's broaden the topic a bit and talk about use cases timeline and general health of the quantum ecosystem. Where are you seeing the biggest impact today with Quantum Computing and what is going to happen in the next couple of years in terms of use cases and the general ecosystem in your opinion?
Florian – Yeah, there's so much. So I would say right now. Quantum Computing already has a strong impact it's because it's it's still a technology that will mature further but it's something that we can use today in a hybrid way. So not limited to Quantum, Quantum chips only but in a hybrid way to efficiently solve problems so that's happening already and that's very interesting. So when you look at the landscape and also you talk to Industry then people would very often think it's still a technology that can only be used for proof of Concepts, that can maybe sometimes be used for creating intellectual property. So you develop some small scale algorithm that for example simulates a small molecule but then you know over time in the future more powerful systems will be able to simulate industry relevant molecules with the very same algorithm but when you look at it differently and say Quantum Computing is an evolution of high performance Computing and you just take everything you have already and you combine that efficiently then it's definitely happening today. So of course it's more than bringing these two together. It's a lot of work that needs to go into operating systems. One of which we developed by the way but operating systems in general are important. So when you look at how a business software developer uses their computers today to develop software they really have to worry about how to address memory or how to address segments on a chip because the operating system does that for them. The libraries on top the programming language that does that for them and that's where we want to go with Quantum Computing and that's what we all have to invest more but in terms of use cases I would say everything optimization today already is high potential not necessarily we can implement the supremacy use case, the industrial supremacy use case where you think you do something with a hybrid system or a quantum computer that a classical, purely classical system can never solve but still when we look at distributing parts traffic optimization logistics today already with a hybrid approach can outperform best in class or even sort of best in business or even best in class algorithms. So that's something we can do already. Second, I would say machine learning. So every time you think about machine learning you have two things. So one is how accurately does my system learn from data and the second thing is how long does it take to really train so and that's two of the big hurdles. So the second one not so much I'd say but still so if I'm faster say I'm talking about self-driving vehicles and you have environment perception happening and training these algorithms. It can take maybe a week, maybe two weeks if you do it from scratch, even with hundreds of thousands of cores. What if I can do that in a day instead or maybe in an hour, that would drastically improve my innovation cycles. I can redeploy these algorithms a lot faster and be faster than my competitors and the second thing of course and most importantly is accuracy. So if I train on a machine learning approach no matter what the problem is in a hybrid way, can I outperform a classical system? Can I be better? Can I obtain more accuracy while still keeping generality? So that's always a thing in machine learning. You want to be able; they want the algorithm to be able to identify as many cases as possible even edge cases but also be very accurate in everything that it does and if I can do that better with a quantum computer, which we it's a in a hybrid way which we can do already today, that makes sense. So for all of these use cases I think it's already happening but more in the future when I think about medicine or Pharmaceuticals or you want to develop new drugs and for example simulate quantum mechanically how a protein is affected by a drug then I mean we can go completely crazy here and think about personalized medicines or develop medicine really specifically to Floria for example to what I have so even if I share the same characteristics of a disease with another person our treatment may be completely different because our bodies are different they're not identical. That's not how we treat today. So I think these use cases last all the other material science or simulation use cases. They will really do a lot of good for the people. So that's why I'm very excited already to think about it!
Rene – Yeah, and again I cannot agree more because like all the things you mentioned we also seeing and especially in Quantum and smart optimization is already providing quite some good benchmarks already and also Quantum machine learning, I had a guest in the previous episode also talking about some qml use-cases and it's fascinating how you even with these small quantum computers we have today you can build these hybrid neural networks where one of the lateral layers basically is represented with a Quantum gates and so on and you can represent multiple states at the same time and that's just perfect and much closer to the real deal than our digital information, right?
Florian - Yeah it is. You should have told me before that you're deep in the field already but that's exactly what it is. Yeah, it's exciting so we achieve higher representational power by integrating a Quantum circuit and purely classical machine learning models such as neural networks and that is or the excitement begins because that means I can use less computational resources but also at the same time be more accurate and ideally also faster.
Rene - Yeah and it's very demanding not demanding but it's very in time to have this available because like you're saying right it's like these models are getting bigger and bigger like GPT3 what cost it like 10 million dollars or so to train it and they just announced GPT4 which is magnitudes even bigger and has even more training data built in and it will catch just a few more factors multi-train and if we can do this more efficient and faster with qml. Well, this is exciting also from the sustainability perspective of course. I'm sure we could talk for many more hours but I'm I've also fully agreed like this technology will help us to overcome and to solve humanity's toughest problems that we can, of course, solve with technology right, that everything we can solve with technology as humans but like also climate modeling, you know, all that kind of like super complex high-performance demanding stuff which is unsolvable with today's silicon-based computing. It has this potential and also drug like you say drug discover protein folding all of that stuff like so much potential, great personalized medicine, that's great actually.
Florian - It is very exciting, yeah and I always think about the good the technology can do so of course, there will always be malicious actors using it for the bad stuff but if you think about the good that technology can bring, if we applied the correct way, that's tremendous.
Rene - Exciting time we live in and like I said we could talk for many more hours, but let's keep it here. let's keep it at 20 minutes around something like this. Thank you so much, Florian, for joining us today and sharing your insights. This was very good and very much appreciated.
Florian - Thank you very much for having me. Thanks for the great questions and discussion.
Rene- Well and thanks for yet another episode of QuBites, your bite-sized pieces of quantum computing. Watch our blog and follow our social media channels to hear all about the next episodes. Of course, subscribe to our YouTube to not miss any new episodes and you can always watch the previous episodes on our YouTube as well as on the website with a full transcript. Well again, thank you and take care and see you soon! Bye!
