QuBites 2.3- Airbus Quantum Computing Challenge

Valorem's QuBites video series breaks down Quantum Computing concepts and use cases to help business leaders learn more about the next wave of technology disruption in quick and easy to consume episodes. Alongside other experts in the field, our Global Innovation Research & Incubation Director Rene Schulte shares expert tips on how to integrate this emerging tech in your business to improve operational efficiencies and drive business outcomes. In the 3rd episode of QuBites Season 2, Rene welcomes Machine Learning Reply Engineer, Giovanni Pilon to discuss his team's winning submission for the Airbus Quantum Computing Challenge.

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 the Airbus Quantum Computing Challenge and for this I'm honored to have no one less than the leader of the winning team Giovanni Pilon. Ciao Giovanni and welcome to the show! First of all, congrats for winning this exciting Airbus Quantum Computing challenge. How are you today?

Giovanni: Hi Rene! First of all, thank you! Glad to be here to talk about quantum. And yeah, all good!

Rene: Awesome. Before we start, can you tell us a little bit about yourself and your background as it relates to quantum computing?

Giovanni: Yeah, so I'm an engineer and I've been working for Machine Learning Reply for the past three years. My background is a mix of mathematics and computer science, which is pretty much what you need to deal with quantum computing. And I've been doing research on this topic pretty much since I joined Reply.

Rene: Nice. So, tell us a little bit. I mean, I already gave it away that you guys won the Airbus Quantum Computing Challenge and that is super exciting of course, but first of all what was the Airbus Quantum Computing Challenge all about?

Giovanni: Yeah, so the Airbus Quantum Computing Challenge was an international challenge held by Airbus through 2019 and 2020. And their goal for this challenge was to explore how quantum computing can be used to solve some real business problems that they have in the aircraft and aerospace industry. And this is what's really interesting about this initiative, in my opinion, which is the fact that these are real business problems, it's real tasks that are difficult to solve before [with] classical algorithms. And so, they wanted to explore, at least from a theoretical point of view, how they can benefit from this new technology and how it can be applied to these business problems.

Rene: Well, that's also the goal of our show here with QuBites, focusing on the impact we're already seeing today with applied quantum computing. Because a lot of people think basically that this is still science fiction [and it will be] many many years out until we have true quantum computing hardware with quantum supremacy. But this is not the case. We can already apply quantum inspired computing, quantum algorithms, etc. even on classical hardware and achieve great results, especially in the optimization space.

But let's dive into your [Airbus Quantum Computing Challenge] project and what your team worked [on]. So, maybe you just want to tell us a little bit about your team and also what was your project about and how did you manage to win it of course?

Giovanni: Yeah, yeah. So the project was focused on a problem [related to] aircraft loading optimization. And what you have to do in this problem is to load as much weight as possible on the plane while respecting a lot of like constraints. So, for example, you shouldn't load too much weight. You have a weight maximum/weight limit. Then the center of gravity should stay between an upper bound and a lower bound, otherwise the plane will not be balanced. And also, the distribution of the weight should stay below a distribution curve that they give you as an input. So you have a lot of constraints to respect and these of course are meaningful constraints from a flight perspective. My team was made [up of] me and two colleagues of mine from Machine Learning Reply. But we have to say that Reply was doing a great job and is still doing a great job of quantum. In fact, we had a quantum computing hackathon back in 2019 and I think we had multiple submissions, like three or four, for this challenge. And speaking [for] our [Airbus Challenge] team, what I can tell you is that we had a lot of focus on the mathematical correctness of the model. We were able to solve all the constraints. So, all these different constraints we were able to map in the same quantum computing model and we were able to solve. So, proving that [the model] was correct. We were able to batch party both on real quantum monitors, because as you said we are using quantum optimization, and also on classical servers. And another interesting thing about classical solver, quantum-inspired solver is that they’re useful not only to have a solution today but also, in this case, to prove that the mathematics and the equation where correct and that in fact we were getting a feasible solution. And another thing I want to mention is that it's not only about having a solution today but also about being ready for this technology if we keep getting more scaling, more qubits in the future. So, I felt that the interest of Airbus was somehow also theoretical, meaning what is the model, what is the math, that [will] allow us to solve this problem of these machines in the future?

Rene: Gotcha, well that's exciting! You're basically saying you developed a mathematical model or an algorithm/quantum algorithm that can then solve this big challenge of how can I optimally load my plane so that I maximize the output, right? That I have the maximum amount I can transport while also not having too much fuel and also, of course, fulfilling all the constraints, right? [The plane] needs to go from [point] A to B and maybe in between a few more stops. So definitely a big challenge and I mean when you tell me you actually solved it with a model you guys developed, that is not just a theory but you also were able to run it with a server, to benchmark it, that is exciting! Because it's not just theory then, it's actually real, practical and you have proven that it works. Well, awesome, again congrats! I mean it's for good reason you guys won that for sure! This is, like I said, it's a big challenge. So what are you guys doing with this now? What are the next steps? And what is Reply doing in these areas? And can you maybe share some examples of other current impacts you're seeing?

Giovanni: Yeah, so of course we are always staying updated. You know, in this field, you have to stay continuously updated on the latest trends and achievements. So that's for sure one activity that we're constantly doing. And also, working as I was saying with the classical solver like CPU or GPU, we are able to do simulation or let’s say provide a quantum inspired solution for this type of model. And for example, one use case that I can share, something completely different from the Airbus one, is Porfolio Optimization. It’s something we developed a couple of years ago and also in this case, the tricky part is that this kind of optimization, like in the one of the aircraft, are combinatorial problems usually. So once again for classical algorithms, it becomes really difficult to test all the possible combinations. And in this case, we're still expanding and extending the work that we've been doing in the past and we have some prospects. So that's another use case that’s interesting for sure.

Rene: Yes, I can totally see that. For example, how can you also optimize your financial investments, right? Because these are like super complex problems, I mean these are multi-dimensional optimization problems. And you need to find your way through all these dimensions to find the optimal solution.

Giovanni: Exactly, exactly. You can imagine exactly like that. If you think about the combinatorial problem, it's a lot of combinations of choices and the amount of combinations that you have to test just scales exponentially. So you need an alternative approach, either classical or quantum.

Rene: Last question, the Airbus Quantum Computing Challenge, was the prize, what did you guys win and how many other teams were actually also taking part in the contest?

Giovanni: Yeah. So, the prize was a collaboration with Airbus. So that's something that's ongoing and that's it pretty much.

Rene: How many other teams were there?

Giovanni: So, at the beginning there were something like 800 people registered for the challenge at the international level. however, the amount of valid submissions, so that’s the submissions that were actually accepted and evaluated, was just 36. So apparently providing a submission was not an easy task.

Rene: Yeah, quite a big filter. But anyway, I mean…

Giovanni: Our paper is online, sorry, by the way. If you're interested, it's on our site now. We published it just recently.

Rene: Oh! That's good to know. We will definitely put a link into the show notes here [so] that people can access that and check out and read all the details about that paper. But, I just wanted to finally say it's pretty amazing actually if you think about 800 submissions initially, and you guys won out of that. Although they reduced it to 36, but that is a huge achievement, so again congrats! Unfortunately, we’re already at the end of the show. Thank you so much Giovanni for joining us today and sharing all about this new exciting work you are doing and the Airbus Quantum Computing Challenge, very much appreciated. Thank you, Giovanni.

Giovanni: Thanks Rene!

Rene: Well alright folks, thanks everyone for joining us today for another episode of QuBites, your bite sized pieces of quantum computing. You know it: watch our blog, follow our social media channels to hear all about the next episodes once we publish them. And so long, take care, be safe, and see you soon. Bye, bye!