Transcript
Rene: Hi welcome to QuBites, you’re bite size pieces of Quantum Computing.
My name is Rene from Valorem Reply and today we’re going to talk about Quantum-Inspired Optimization and I’m very honored to have a special guest today here with us, our expert Marco [Magagnini from Data Reply]. Ciao Marco, how are you doing?
Marco: Ciao Rene, very good. Thank you very much for inviting me.
Rene: And thank you for sending me over the aperitif kits, I got a little nice package with this great card in here and I got a bunch of well, aperitifs, so thanks for sending it over, cheers my friend!
Marco: Cheers!
Rene: Can you tell us a little bit about yourself and your background as it relates to Quantum Computing?
Marco: Yes. Very briefly, I got a degree in Physics and Ph.D. in Quantum Engineering. In the last 20 years I’ve worked in consulting for information technology. For the last 3 years I’ve been working on Quantum Computing for Reply.
Rene: Awesome, that’s quite impressive! So, you’re the right expert to tell us what is Quantum-Inspired Optimization?
Marco: Yes. So that is a very, very new field. An applied field in which you use Quantum Algorithms to find the optimal solution [to] a lot of problems. [An easy analogy] to understand [Quantum Computing] is when you have to pack a lot of dishes in your dishwashing machine. You want to pack a [lot of] dishes, the glasses, etc. It’s not so easy. Your solution [is to just to stuff them inside a dish washer]. However, you can find a better way, a more optimal way. And computers like quantum computers, they are very good in finding those optimal ways [for us]. So, this is a new field. And I hope that this [analogy] explains a bit, this [new] field.
Rene: Well, I can’t wait until we have Quantum Computing dishwashers! I mean, that’s just a challenge that everyone of us has to face every day pretty much, right? That is awesome. What are some of the typical optimization problems you see with other clients?
You typically have an optimization problem, like [your example] packing a dishwasher, or it could be cost savings [problem], where in the end you want to reduce the energy you put into the system and maximize the output, right? You can run that with Quantum-Inspired Optimization. You can actually develop an algorithm that is designed with Quantum Computing in mind, so it can run on Quantum Computers, but it can even run on classical hardware, right? And so, you’re already seeing that these Quantum-Inspired Optimization Algorithms are outpacing your classical state of the art optimization algorithms even on classical hardware, right?
Marco: Yes, we are seeing this kind of scenario where you can run [quantum algorithms] on classical hardware and we are seeing a lot of real-world use cases. You know, [Reply is] among the 5 finalists of the Quantum Airbus Challenge [in which the goal was to use Quantum Computing to maximize space in an] airplane for transporting goods. So, you know there is a lot of these kinds of real-world [applications for Quantum Computing already]. Another project, that we have done [was focused on] optimizing the [journey] that [frontline workers of an energy company have to take] to do maintenance on energy lines. So, [field technicians] have to do a lot of things in different places and you want to [find way to help them] travel less and work more. And these kinds of problems, it seems strange, but they are very very difficult to solve. And you can solve with an approximate way. But if you solve better, you save money, you transport more goods or you travel less meters, you do less pollution, you use less gasoline. So that is the thing. You can run on CPUs, GPUs, QPUs.
Rene: Nice! So you can take these advanced algorithms designed for Quantum Computing, run it on classical hardware like an array of GPUs, which are optimized for computing a lot of matrices, so you can take advantage of that and run that. It’s pretty amazing, and one of the problems you described sounds like the classical traveling salesman problem, which is this complex problem in computer science. And I know that one algorithm is gaining a lot of attention these days, it’s called QUBO, what is it and why is better than classical optimization algorithms?
Marco: QUBO it stands for Quadratic Unconstrained Binary Optimization. At the end of the day, you model your problem in a quadratic way. So, this modeling is much better, more flexible and more adaptable to real-world problems instead of linear programming. So, you have a lot of real-world problems that are better modeled by QUBO, by quadratic. And moreover, with this modeling you can run [much] faster on GPUs, because GPUs do matrix multiplication and [with] quadratic models you have a lot of matrix multiplications. We have studied this in depth in the last three years. We have published a scientific paper on Springer, Quantum Machine Intelligence [with] a peer review journal and we have done 5 projects and we have seen that this is the real world. So, we are in production [and] we are delivering value to customers through this kind of approach.
Rene: That is awesome! And in the next episode we will actually talk much more about the impact you’re already seeing today with some client projects. I think that’s great to hear about Quantum-Inspired Optimization, QUBO. I think we all understand now what it is. And so, thanks for joining us, we’re already at the end of the show. Thanks so much for providing your insights and joining us today, very much appreciate it. And thanks everyone for joining us for another episode of QuBites. Watch our blog, follow our social media channels to hear all about the next episodes. Take care, be safe and see you soon, bye bye!
Marco: Thank you very much everybody, bye bye.
