Quantum Programming: Classiq + Wolfram

Right hand like you know, doing everything like you know, that we do in academic innovation. So she'll be the one starting like in our lives.

Sam. So you ready? I'm going to make the cue Classic cue ready.

All right.

We are live. Go ahead Yu.

Hi everyone. Welcome to join our academic innovation support live stream. Today we're back to our Quantum topic and we will have a special conversation with Classic Q. So joining us is Ralph, the quantum application engineer from Classic Q and also Eric, the head of technical marketing. Of course, one of our regular guests is the lead of Warframe Quantum team Mats. So I will let them start conversation.

Excellent. Thank you. Li so very happy to have Ron and Eric, two of our good friends and partners on Quantum. Let us start like you know, with this big confusion about like you know, the name of the companies, like you know, having, you know, Q. Because all people when they read that they would say shall I pronounce it as classic or shall I go Classic Q? Like you know, how should I pronounce the name of the company? Let's just start with that.

Yeah, we tend to go with classic ourselves, but you know, a lot of people also call Classique, you know, it's a little fancier. But no, we tend to call ourselves classic.

Nice, Nice. So let's see like you know, our plan for today is to have a conversation regarding like, you know, the great functionalities that you guys have been developing on Quantum spatially, like you know, considering like, you know, real world applications and of course like, you know, integration with Wolfram Mathematica that opens up the possibility like you know, especially in academia for students, you know, users to explore more like, you know, great functionalities that you guys have. So Eric, Ron, whoever wants to take the lead on that, let's, let's, let's start that conversation.

Yeah, absolutely. So I mean again, thanks for having us on here. I've been loving working with, with you and the rest of the team and let's show a little bit of a slideshow. I got to kind of frame things out and yeah, if anybody has questions during the the live stream, feel free to ask anytime along the way. We, I, I talk about Quantum all day, every day and it's what I love to do. So feel free. All right. So yeah, well, we're classic, we're quantum computing software and as Mod said, we're working on integrating together with all the great things happening with Wolfram. So what Classic is, is we're an end to end software development platform. We have a programming framework that has high Level modeling. We're hardware agnostic, which means any of the code you build will work on any of the hardware you want to run on. And it's for developers with both, you know, high quantum knowledge and for those who aren't coming from a quantum background as well. And you know, for, for those who aren't as familiar with quantum. Right. Quantum is just a new way of doing computing. Instead of bits, we're using something called qubits, probabilistic bits at the end of the day. And there's some pretty cool things you can do in mass with it. But the reason people are interested in the first place, from a big picture point of view, it's going to allow certain things to be done essentially less computationally intensive, certain calculations. So the same way that the advent of the GPU made certain types of calculations less computationally intensive to perform and it enabled a lot of cool things, we were able to do things faster. Same thing's going to happen with quantum computing and quantum development. And there's lots of different algorithms that are out there in the universe of quantum and that means there's also lots of different applications depending on what you're looking for using it. So there's a whole myriad, you know, whole sleuths of problems and we'll get into some of them later. But it's a wide reaching technology that's going to affect a lot of different industries. And right now just so kind of framing things out, the state of quantum computing. A lot of people call this the, the NISC era. It's a word you're going to hear if you, you know, do any googling of quantum noisy intermediate, quantum, you know, the hardware is a little bit noisy right now, but you know, there's lots of different ways to make hardware, lots of different kind of hardware modalities, the physics of making these kinds of quantum particles. Lots of different manufacturers and developers are building out different machines, but they're all increasing in size essentially. Right. That's kind of the message. There's lots of different ways to do it. We're in this kind of noisy era because the machines are being built out and developed, but they're all scaling quite lovely. And frankly I haven't been part of a bleeding edge industry that's actually kept to their roadmap. So I'm actually quite amazed that we actually have kept to each of these hardware roadmap for companies and kind of the modality for programming these systems is at this kind of assembly level language. If you're looking at this Code what quantum coding looks like. Each of these lines is a qubit or the particle that you'd be kind of doing your computation on. And each of these blocks on the screen are the, essentially the calculation of the, the operation you're telling that specific qubit to do. It'd be the essentially the equivalent of Boolean logic for, you know, for your bits. And kind of this is what we hit with the, what we call the quantum software challenge, is that it's kind of like punch card programming, this kind of modality of programming and large, in part, this is not going to scale. I don't like to use the word impossible to design very large quantum circuits, but I'll say grossly impractical. I think anybody, if they spend a long enough time, someone will prove you wrong and say, oh, I can actually do 125 cubic quantum circuit by hand. I'm sure you could. How, I don't know how long that would take you. So grossly impractical is the, the verbiage I, I tend to go with. And I don't think anybody would argue against that. You know, doing it by hand, that would just take you a hot minute to build any kind of large quantum program. And the same way with classical programming, right? If you're trying to build an application that's worth its salt, it's not going to be on five bits, right? It's going to be thousands of bits, you know, and you can't be doing that in Boolean logic. It's just grossly impractical. Yeah. And kind of programming for these systems needs kind of a holistic approach at the end of the day, right? You have your problem space, you have your solutions you can develop, but you need context around that problem. What is the hardware you're going to run on? And there's so much kind of minutiae when it comes to programming quantum stuff right now that you have to know what hardware are you going to run on? What's the connectivity for that hardware? What are the native operations for that hardware? You know, how do you balance the efficiency of what you're trying to code in versus how that affects the hardware and the algorithm you're putting it into? And largely what quantum and sorry, what the classic does for quantum is it abstracts out a lot of that context and what you'd have to know as a programmer to build efficient quantum circuits so that you have to be just an expert on the problem you're trying to solve. And you still have to be quantum to solve it, but you don't have to be kind of that low level machine quantum expert to program these systems. Which is why Classic helps out for both quantum and non quantum experts. It's more on the programming end. You program efficient quantum code that uses the hardware instead of having to know all the minutiae of the hardware itself. So it abstracts out all those kind of hardware constraints and it looks through a huge phase space of different possibilities for you to get that efficient quantum code that will run on whichever of the hardwares you're deciding to choose and go on. And hopefully what we're trying to do with you and your team's mods is kind of help people go through the whole quantum computing roadmap. Right. A lot of people start off with kind of the exploration phase and they built some small quantum circuits and they stop. They don't kind of go on to the next stages. Right. And that's what we're really hoping to do with Wolfram and everything there and all the great applications you have is I'm heck, I've done this both in the lab and in industry. When I worked for Kodak Alaris, actually I was in their innovation department for a while and we used Wolfram. And you start off with a small program, a small project, then you gotta pilot it somewhere. You gotta build some, some infrastructure. You take that kind of rough work you did and you scale it up. And that's where a lot of clients and a lot of people working in quantum in any, you know, scale phase, it's that scale up that they have a lot of trouble with. And how do they investigate this in a way that is actually practical? Yeah, and I can go into some more things here in general, but there's a lot of reasons people are investigating it. A lot of, you know, McKinsey put out a report saying there's a lot of money coming into quantum in the near term for a few kind of key industries. There's some key algorithms that are really important. And depending on how the conversation goes with the audience, I know we can geek out about some of these variational quantum algorithms and things like this kind of the near term quantum technologies. But I'll kick it off back to Ron, who can talk more about what we're doing on the integration and how we're working together.

Excellent, excellent. So basically, like, you know, one thing that, like, you know, Ron, you can ask, you know, I'm, you know, sharing a few thoughts with our users. You can go ahead and share your screen. One thing, you know, regarding, you know, Mathematica that many of our users like you know, they no less compared to many other functionalities. And Ron like you know, he's going to showcase like you know, a few of them is many of like you know, these platforms to do fancy quantum computation such as what we have in Classic. They are built in different, you know, ecosystem. Like you know, they are building, you know, Python environments. And like, you know, some of our users they would say gosh, like you know, this is Python and like I'm a mathematical user, it's like, you know, what about the integration? And this is the part that I would like to remind the users that like you know, the you know, encapsulation of you know, other programming languages in Mathematical Notebook it can be done so easy. And if you do that like you know, you are going to benefit from the power of both platforms like you know, you can get all the functionalities that you have in Classic plus any other functionalities that you can use like you know, in Wolfram Mathematica. So let's look at a few of them. Ron.

Yeah, great, thank you so much Meds. And thank you Eric for the great classic introduction. So just as a small agenda, I will present two examples. The first one will focus on what can we do with arithmetic in Classic based on like a use case of Brownian motion which I will show you. And the second one will be more in the chemistry field finding optimal solution to molecules in a classical way versus a quantum way and and comparing them. And like I'm open to discussion, so feel free to stop me however you like. Okay, so, So modeling Brownian motion. So let's start with what is Brownian motion anyway? For who doesn't know it is that particles in liquid or anything are moving randomly up and down because some other particles move them. And like the GIF showing it the best way that there is some random behavior of the movement of this particle. And there is like a lot of analysis that is done on by in the physics departments everywhere to model this Brownian motion. And this is in physics taking it to the financial world. If you think about a market where people are buying and selling some assets. So it's behaving very much like the particles that is pushing up and down. Any exponent of this up and downs represent the actual price of some assets. So which has also some random behavior. And there is a, there are mathematical models behind such behavior which are bound in motion actually. So let's see how can quantum computer help us to analyze such behaviors in financial markets or in physics of Brownian motions. So Eric gave a great introduction and told you that the difference between a classical bit which can be 01 to a quantum bit which is can be 0 and 1 at the same time. So the idea is that we brought. One of the ideas about with classic is that if we have quantum bit, why not having a quantum number? So if we I can be 0 and 1 at the same time. So a quantum number with 4 qubits can be 16 possible values at the same time. So by using just n qubits I can hold two to the power of n values at the same time. It's coming to imply about the power that quantum computers bring. So if it takes thousands of qubits, so you have one and 300 zeros later of numbers that you can hold in a computer at the time. So. So we have a quantum number. And now we would like to see how the idea of compiler that is able to to tolerate quantum numbers and make and operations that feels to us basics in the classical world, like additional multiplication is done in the quantum world. So now let's be more practical connected to some use case. So J.P. morgan and Chase bank, the biggest bang in the world, maybe they published an article on February that is telling how to model option pricing in financial market using Brownian motion. The article was published in February, it was totally theoretical and it says that the Brownian motion is some random variable that is described with a lot of summation and multiplication and a polynomial which is defined in a recursive way. The Chebyshev polynomial, if you know and that something that writing in a classical computer is trivial because in one line you could do multiplication and addition and everything is okay because we're regular to the fact that the classical computers are already giving us those basic, basic elements. So this was totally theoretical article and there is a good reason why. Because if you think about the actual implementation, a lot of like the gate level operations that are done in order just to make a multiplication or addition at the basis, at the basic level, it's a lot of gates. Just as in a classical computer, when you multiply A by B, there are a lot of operation hidden there. So using the classic platform, which gives a high level programming differently from other languages which are still like assembly, as Eric mentioned, make it impractical to implement such operations like arithmetic. It's telling us that we can maybe actually take this theoretical article and implement it. And two weeks after it was published, I took the classic QMOD language and I wrote the developments and I was able to Create a big circuit and I will show you how it was done inside the Mathematica notebook. Okay, so now first time that we see how like implying of how integration with Wolfram can be powerful. We still have more work to do in terms of integration and of course we need some bigger quantum computers that we can actually run the circuits, but we can create big circuits easily with this combination of Wolf form and Classic. So for example, let's say that I analyze classically analyze with Wolfram Mathematica Born and motion modeling there is like a built in function of the winner process, which is Brownian motion. And I can look at the goings, ups and downs of some asset or the exponent of it, which is implying that an asset will go down if it's supposed to go down and low. And of course this is random and we can run again and look at different occurrences of this behavior. So and of course we can do some more smarter things inside Wolfram Mathematica. But now let's say that the user would like to investigate the Boundian motion with the computer. So we can import Classic, install the package import Classic and let's see how can we implement this article using Classic. So we saw this one and we say we want to implement quantumly to make a quantum multiplication of these formulas and quantum summation. By the way, quantum multiplication and summation means that in one operation for addition, for example, if I have some quantum registers A and B, which is in a superposition of many values. So for each of the cases, in each of the worlds that the variable value exists in, I have to do that. So this is the actual operation for addition. But in Classic it will look almost trivial. It is not in another languages, but as you can see, let's take, let's start with this multiplication and addition. So with Classic you do it with just one line. So this multiplication is actually a big quantum circuit. In this entire summation and multiplication, it's a big quantum circuit. And I will show it to you.

But it's now like Ron, like let's provide like, you know, some context for our audience. Like, you know, you mentioned, you know, QMod, which is, you know, the programming language that you guys have been developing, like you know, specifically on quantum. So elaborate a bit more on that. Like, you know, what is it like, you know, when you talk about qmode, like, you know, what we are talking about, like, you know, and how users should think about that.

Yes, so, and maybe it's a great question. It will make order of how the language actually Looks like if you start with. Let's take the entry point like where everything starts. So the actual variables as you see, I told you before that we have qnam. So there is some variable. Its name is T, the type is qn. So behind the scenes I will show you how the circuits looks like. So behind the scenes you can see here that T is a quantum variable. In this case it just. I chose just one qubit, but it could be more. But it represent some variable. Once I have a pointer to this variable. So I can start and make some operation based on on this variable. For example, I can do Adama transform, which taking T which starts from a state, no superposition, no quantum at all. And once I apply H, I can be in a superposition. And so this is like the most basic thing. More complicated variables are now taking these variables which are in superposition and we and creating even bigger superpositions and enlarging it more and more. And how it is done in operations that are taking quantum variables. Here are my variables that I need for this summation, multiplication, taking the variables and making the multiplication on them and addition. And let's see how it looks like. So if I zoom in and I look inside so I see that using the platform, I can take specific variables and make multiplication on them and make subtraction, etc. Etc. And then what I actually get is the time if I start with some superposition of values of some variables. So I can create new variables with arithmetic operations based on the the other one.

Now let's like scroll down, let's look at the final result and like, you know, because you know, some of our audience, they ask a few questions, let's look at the final result. And then based on that, I'm going to ask you like, you know, some questions. Okay, so here you're adding the constraints.

Yeah. So what I would like to mention about, I talked about the language and I show you how you can take theoretical idea and put it into code. Now the another thing that the product is the compiler is doing is running is solving optimization problem. Its target is to choose the most optimized implementation for what we wanted behind header. You can implement other in different ways. But some ways are better in some manners and some are better in another manners. So for example, if I want to compile for Atom computer, I will compile it, I will get some result and when I will compile it for Ion's computer, I will get another Result. So like two different implementations.

Got it?

This very big circuit with 130. I don't know how much qubits. So you can see in the data here that I get different results for different types of computer. I could not execute this example because there are too many qubits and the hardware is too noisy and the simulators are too small. But it's just showing you how easy it is with the language to create big circuits. And in the next example, I will show execution.

Okay, so there are a few questions from our audience, but before jumping to that, we have another guest from Classic joining us, Shai Lev, who is VP of Strategic partnership at Classic. Hi, Shaik, welcome to our livestream.

Hi, Mads. And thank you for hosting us for this really exciting session.

Sure thing, sure thing. One of the interesting questions that I guess you helped me with that is our users, they just jump ahead and they ask about what's the joint vision of Classic and Quantum for future of quantum computing. So if you would like to elaborate more on that and the vision that we have in our partnership.

Yeah, so I think that the vision that we all see in front of our eyes is basically creating the new standard for quantum computing software. And by combining the capabilities of what you guys bring from Wolfram together with the deep expertise that we have in quantum software, I think that this is something which is really achievable. You know, at the end of the day, creating a quantum computing software is something very wide and very heavy. And if we take a comparison from the classical world, you can't really see any one company doing all of that by itself. I think that joining forces for us, providing this really basic, I would say even operating system for quantum software, together with your expertise and together with your knowledge, that's something which is really, really unique.

Thank you. Thank you for your input. And this is an invitation for Classic and Wolfram users in academia, industry and government. Like, you know, we would love to hear from you guys, like, you know, the real world challenges that you have on applications on bleeding edge of like, you know, science and technology. And we would love to help, like, you know, both companies, they have very unique expertise and like, you know, and I'm sure, like, you know, it's gonna open up like lots of new possibilities in the world of Quantum. Now there is another technical questions I'm gonna ask it like, you know, Eric. So I appreciate if you're pointing that like, you know, our users, like, you know, they just jump to very technical question. One of our users, they ask about like, you know, the future of, you know, quantum classical hybrid approaches especially, they ask like you know, how like, you know, we think about, you know, the future of high performing computing and HPC facilities.

Yeah, now that's a great question, right? I was, I go to a lot of these trade shows and events at supercomputing and whatnot. And the vision that the industry has, and not just classic on kind of where quantum computing will be at the end of the day is part of that HPC stack, right? It's going to be part of all these high performance computing systems. In the same way that in those Systems you have CPUs and you have GPUs and you have TPU tensor processing units. You'll also have at the end of the day a qpu. The same way you send the right kind of data to be processed on the right processor, that's the least intensive for that data to be processed by. You do the same for quantum, right? You wouldn't send all of your graphics to be handled by your tpu. It's less intensive to run it on the gpu. So we send it to that. In the same way there's going to be types of problems that are much less resource intensive to run on the qpu and we'll use it for that. And it's going to be this holistic approach at the end of the day kind of all of these processing units working in, in chorus and in harmony.

Excellent. And this is like, you know, one of the core areas that like, you know, both, you know, Classic and Wolfram, we are actively engaged with our national labs here in US and you know, other places because as Eric mentioned, one of the key questions is when you are bringing quantum value into these computations where it can help the best. And this is basically like, you know, one of uncharted territories that like, you know, that I think like, you know, it is an invitation for many researchers and of course people from the industry that there are very interesting promises from quantum computing that, you know, the term that we usually use is quantum advantage. But you should keep that in mind that like, you know, it's not, it's a completely non trivial question to say how quantum value emerges out of your computation. And it needs like, you know, as mentioned, like that's why we need to join forces because these questions are difficult. But you know, both companies, we are actively working on that and you know, we believe we have some innovative ideas to tackle that. So now I'm going to go back again like, you know, to Ron. And Ron, like, you know, you had, I know you had some more interesting specific examples on Chemistry applications. And by the way, for our audience, we know that we are choosing very simple, like, you know, cases. It's mostly just to show you like, you know, what's possible and see it as an invitation for you guys to explore more and help us with innovation.

Yes, we can see your screen. Great, thank you, Mads. And someone asked about the hybrid execution. So it was a great question just before this example, because here we demonstrate the variational quantum eigen solver vqe, which is actually a hybrid algorithm, and how it's done and how do we leverage the advantages of Wolfram and Classic in this notebook. So, generally speaking, one of the interesting problems which are very important for chemistry applications is finding the minimal energy location for atoms in a molecule. And it's helped them to develop molecules and important materials like fertilizers, for example, in a way which is more efficient, in a way that is less computation intensive. And the VQ is an example here it's just an H2 molecule, a small example, and it shows how can we use Wolfram and Classic for that. So I will just show what I would like. I will begin with what I would like to show. So in order to olve, using classic QMod, the location with the minimal energy of a specific molecule, we use the VQE algorithm. And the VQE algorithm is taking a molecule and running on different combinations and locations. And it will give the minimal energy using Classic, you can define the molecule in a way that is relatively simple. You just have to mention what is the type of the atoms that you are interested and the location here, which is a variable, because I would like to change it every time and find the best location. And then everything behind the Miltonian construction, all the ugly words that you need for VQE is done using the compiler and we run a hybrid algorithm execution of it using the scene main. It will look a bit more beautiful in the future. It will add just small integration problems with wolfarm. But you can see here we have four lines. Please run a VQE algorithm, this quantum circuit and convert the molecule result accordingly and give me the actual solution for the for the molecule. So now once we finish to run the algorithm, I ran it before the live stream, so we can see that in different location of the of the hydrogen atom there are different energies. And we are looking for the location which the energy is minimal. And we would like to use the Wolfram notebook in order to compare between the result that we got from the VQE quantum algorithm and the actual classical result from Wolf A notebook.

So Ron, you know, if you scroll up like, you know, into that graph, let me just like, you know, comment on one thing, like, you know, the dash line that, you know, that our audience, like, you know, especially those coming from academia, that's the simplest potential energy surface that you guys may have seen in your physics or chemistry textbook. And what we call as a molecule, in this case hydrogen molecule, it lives at the bottom of that well. So basically, like, you know, that's the most stable, like, you know, place. And the molecule leaves that there at this potential energy, the simplest one. Like, you know, it has also another term, they call it like, you know, Morse potential energy in the literature. And you see, this is obtained using like, you know, a fully quantum approach. Like, you know, it's completely by quantum, I mean, quantum computing approach, not like, you know, the conventional approach that we have like, you know, in computational chemistry.

Yeah, great, thanks, Mads. And I will show you right now what is the solution from the classical world, which of course I will do with the Wolfram. So with Wolfram, all I have to do is to define the molecule and I'm asking, asking Wolfham to give me the actual distance between the atoms, which is with minimal angios. So Wolfram is telling me the classical solution. What we tell you is the optimal solution is 0.54 angstream. And this is the solution. And now we would like to check if the solution of VQE using Classic was the same. Of course that it's depend on the steps that I chose. I chose if it was dense enough so it will affect the resolution. But we would like to see if this number is close enough to the actual VQE solution with Classic. And we can visualize again using Wolfram, we can visualize the molecules and see that for different distances there are different energies. So we see here that the energy is minimal 0.8. So this is telling me using VQE that the location, the distance between the atoms at 0.8 is really very close to the classical solution, which tells me that this is true. And we can see it also here in, in this nice graph. And so it was of course a very simple example of just to hydrogens molecules. But if you go to the world of fertilizers, for example, like very big molecules are very important applications for finding such optimal solution. Those are solutions that are still not known to the classical because the running times is too big. And we have hopes that this will bring developments of material, of medicines. Anything is in the chemistry world.

Excellent. And by the way, for your audience. Yes, we will share like, you know, these notebooks, like, you know, for sure. Like, you know, you can use them and you can, you know, replicate the results. There are a couple of other interesting questions, like, you know, some of them, like, very technical questions are users, like, you know, the user love to ask technical questions. So talking about, like, you know, maybe like, you know, I'm going to ask, you know, one of these questions from Shai. So one of our users, they are asking, like, you know, they are interested to collaborate with Wolfram plus Classic Team. Like, you know, how they should contact us. Like, you know, that's basically the detailed question that they're asking.

Great. So I saw the question and definitely reach out to us either via Quantum at Wolfram or via ronetclassic IO. And I think that, you know, for us, we understand that we're just at the beginning of the, of this journey. And definitely what we're. One of the things that we're looking here together, both Wolfram and us, is to know the users better, to know what we can do together in a wider scope of work. Because at the end of the day, it will take a village. It won't be done only by Wolfram and Classic together. We need to see what is out there, what is interesting people. And that's definitely, definitely something that we would love to explore together. So feel free to reach out. And, you know, Mads, maybe one point about the HPC center and the integration, I think that we talked nicely about the technical aspect. I think that one of the interesting things to see is that when these kind of centers, high performance computing centers, start looking at their Quantum strategy and they actually even buy a QPU to be an enhancer, etc. That's only their start. I mean, the work that needs to be done on the integration and on the software level and on the application level is very important and very heavy lifting. And I think that that's one of the places that together, as you mentioned, with national labs here in the United States or elsewhere, we can definitely bring their contribution to take that steps forward. So this is, I think, as you mentioned, one of the most exciting aspects of this collaboration.

Definitely. Like, you know, one more time, like, you know, we are, you know, in a live stream, we share our email contact, you know, Classic PR at like, you know, you know, I just share it with my team. Like, you know, they are going to put it on the live stream. So the users, they can directly reach out to Classic Team, they can directly reach out to Wolfram Team and yes, like, you know, this is the area that like, you know, in near future we will see like, you know, lots of interesting applications. And like, you know, one thing that like, you know, we are really, at least from Wolfram side, like, you know, have been enjoying like, you know, our collaboration and partnership with Classic and is both of us, we care a lot about real world application. So we are very good, you know, 35 years of like, you know, woe from, you know, engagement with academia. Like, you know, we know that we are very good in like, you know, doing original research the way that it is done in academia. But on Quantum, in addition to that, we would love to see real world application. And this is the part that like, you know, the partnership and joining forces with a company such as Classic brings lots of value because they are the one who are focused specifically on the best quantum algorithms and the best for specific applications. Now, again, talking about high performance integration, there is another detailed question. Eric, Ron, whoever wants to appoint on that, our users are asking in terms of software design, like you know, model view, controller, like, where do you see like, you know, the integration or the true value of quantum, like, you know, would it being like, you know, modeling, development, would it be like, you know, somewhere else? Like, you know, that's that they want to hear more details on that.

I go for it, Ron, or do you want me to take that one?

Yeah, go ahead.

Yeah, yeah. I mean, I'll talk a little bit about our language and how we structure it and how that can affect things and play around with things. Right. So the QMOD language that you were seeing Ron demonstrating, it's a declarative programming language, higher level, abstracted above kind of assembly. Right. So it's, it's a, you know, much friendlier to code in and kind of lets you abstract out to any of the lower level coding languages out there and available. Because, you know, every hardware manufacturer has their own, you know, low level coding language. Right. IBM has open chasm with their, their whole qiskit thing and Microsoft has Q and Google has Circ. So programming at a higher level helps you integrate and work with any of these kind of lower levels. So if you're looking at kind of the, the quantum stack, you would say, right? Classic would be a higher level abstracted language that then compiles down to any of these lower level languages that then get interpreted by the, you know, quantum processor. And then that processor, you know, when it hits the software hardware boundary, it needs to make sure that are all the operations actually native to the Hardware it's running on. How do you interpret these kind of machine code operations into the system? Let's see. But try to make sure. Did I hit the question right or was it slightly different than what we were getting at here?

No, that was very helpful. Like, you know, Ron, like, you know, would you like to add more on that?

Can you elaborate the question? I try to understand.

So basically like, you know, think about like this. Like, you know, so we have like, you know, a great classical algorithm, okay. And we can run it in one of like, you know, these HPC facilities. And now we are thinking about like, you know, quantum. So where in this workflow, like you know, it has many steps, it has like, you know, integration with the hardware, it has like, you know, code design. It's just like, you know, analyzing the data at the very end. Like you know, the data that like you get after like, you know, the machine does all the grinding. So when you think about like, you know, quantum like that, the case of, you know, the finance case, like you know, that you just showed at the beginning why somebody should say, okay, why I should care about quantum, where in this workflow is going to bring a value Because I have a good classical algorithm and I can run it on my HPC facilities.

Yeah, sure. So, so this field of a quantum advantage is very interesting. The most interest is in the area of the computer scientists that explore what is the actual advantage. So there are, there's like families of algorithms and each one of them is giving different advantage. So first of course is like the most non proven advantage, like no proven yet. For quantum machine learning in the variational algorithms and things like that, people trying, there's no proven advantage. People do find them useful and try, try to prove it and bring it, but no proven yet. And then at the next level you have the quadratic advantage. So it means that if something with a classical computers takes you X, so with the quantum computer it will take like the square root of X. So this is a quadratic advantage like the Grover algorithm which is doing searching like Bitcoin mining and things like that. Also the financial algorithms like the one I mentioned before, option option pricing and things like that. So we see quadratic advantage. Quadratic might be maybe not good enough but of course it is very important when you are in financial market and the one which is fastest is the winner and things like that. And the best family of advantage is the exponential advantage which is splitted into three main algorithms. One is physics and chemistry, which is like the most maybe important thing. Discover medicines and biology and everything. So it's very important for physicists then is the cyber security the Shor algorithm which is can crack any RSA encryption. And we see the entire world is changing the encryption because of the coming quantum computer. And then the last family maybe that I have in my head at least for exponential advantage is hhl. And HHL is inverting a matrix in a logarithmic complexity rather than polynomial. So it gives you exponential advantage over classical methods. So in any application where you have to invert a matrix and there are many of them, let's say a differential equation solver for there are applications we are working with our clients it give also exponential advantage. Now maybe to complete the. The answer to the question if he was asking where is the software coming in this pipeline? So in my imagination maybe I will share my screen. It will maybe help to give better view. So just from looking at the classic platform where there is a model which is defining in a high level language something and you can choose it from a big libraries of many examples and then you can synthesize to compile to a specific backend, visualize it and then execute it into some environ in any environment. Maybe it will be hpc, maybe not. Maybe it will be hybrid and maybe not. So we can imagine that we can connect to many different hardwares. So this is currently how it looks like. And then once it's inside an hpc. So I imagine a cooperation of this software is like we call it on Prem is sitting inside the HPC and not online able to do to create the circuits to create the logical circuits. Also able. Maybe we don't do it today we have just POCs to create also the error correction software stack and. And to manage the entire operation based on those examples.

Now like you know, there is another interesting question by another users like you know, which is a good continuation of what you just explained Ron. So why you guys decided to go toward high level programming rather than like you know, very low level for quantum. Like, you know, what's. What's the benefit of going to the high level?

Yes. So maybe the best explanation I just today this morning I answered many people about. About quantum ether which I mentioned before. This is just one example but I think it's emphasized the best. Why do we need it? Let's take a look on answers. So people was very confused of managing and following the exactly the qubits when they. Just a moment, I'm searching.

Sorry, you answered too many questions.

I'm surprised that like you know, Ron is bravely sharing like, you know, the, his whole screen. Like, you know, it's a scary thing to do, so. But he's so brave.

Yeah. So you can see people were very struggling to implement, as Eric said, grossly impractical to make an adder. And following the qubits, what is the fraction? What do the qubits represent? The fraction, what represent the sign? So it's classic. It's just as you do with Python or any other high level, also using the compiler, you don't have to manage and be aware of your specific hardware. You just write what you want to do rather than how you want to do. And then once you compile it, you just choose which hardware you want to run on. So this, I hope this is answering the question of why, what high level?

Yes, and I think that Mads, it also connects to the point that you mentioned before about both companies having such a great collaboration, not only because of we like each other and the nice smiles, but also because we have a similar approach of being very practical. And when you want to be very practical, I think that the high level approach is, you know, is the one that you basically need. If you're looking into the future and you're seeing, okay, we want to get more complex, more sophistication, I see that as a main anchor, you know, in the vision.

Absolutely. I mean, I was even thinking about what Ron was saying earlier about kind of these proven out quantum algorithms that have this exponential speed up, like hhl, things like that. There's so many applications for these things. But when you think about the practical use of these applications, right. HHL can solve linear systems of equations. And if you look at what that can be, that's a lot of physics simulations like computational fluid dynamics, how water flows, how fire flows, turbulence. And this plays into really huge applications. That number of industries are trying to do like digital twin simulations, right. Trying to build these whole real world simulation ecosystems that can mimic the real world and you can kind of prototype your, your environment in real life. And what's stopping that from becoming more of a reality is the computational intensity of, of building that out. And quantum computing is going to be a piece of the puzzle that helps lower the intensity, you know, the computational requirements too.

Definitely. So I'm going to bravely share my screen, like, you know, related to that question of like, you know, high level, like, you know, programming language. This is also like, you know, as many people know, like, you know, that's, that's the core idea behind like, you know, Wolfram language. And mathematical environment. Like, you know, what we try to do is to help users to have an experience that is as close as they can to what we call as computational thinking. For example, like, you know, look at like, you know, let's go through like, you know, this line of codes. Like you know, here a user saying let's even like when you are not even an expert, you would say like I'm creating a quantum state. The amplitudes are random real numbers. I'm creating eight of them and then I'm going to normalize that state. And this is basically the list of amplitudes on that state. And the next level I'm going to say now create a quantum circuit operator and connect with classic. And then from classic, like you know, the bound, like you know, this is the bound that I'm going to set for like you know, the target, you know, probability. And this is basically like, you know, the philosophy that we have like, you know, in both classic, you remember like in our audience, like the example that Ron showed you, like, you know, you could open up like, you know, excel in any other programming language and you can go into the all the details of like you know, how to get it done at that level or you can do it at this level and basically not thinking about like how behind the thing, you know, things are done. And this is basically the strategy of like you know, both companies, like, you know, we are very good creating the products at a high level that is very useful for the users, the end user from the industry. Basically like, you know, we, you know, a common thing that we usually, you know, tell to many people is quantum applications should not be for a user with a PhD in quantum. Okay, this is not the right strategy. When we are training the future, you know, quantum workforces, we shouldn't think about, oh, all of them. They should have a like, you know, PhD or master in quantum in order to be able to use like, you know, these interesting functionalities. Like even somebody who is measure in humanities and they want to do an optimization, but they know that like you know, okay, I would love to use like, you know, quantum advantage for that optimization. They should be able to run those algorithms. And this is in a sense the philosophy behind this high level programming that we follow in wolf form research. And also classic. So we are almost like, you know, on time, like you know, you know, the last few minutes, you know, our great guests, like you know, Shai, Eric, Ron, Any final words that you would like to share with our users?

I would be happy to say that like being, being Practical. Just before we finish, if our audience feels that they are now very passionate, they, they want to try it by themselves. Just open a Wolfa Mathematica notebook and you can take from the big library. We have also the Git library, just writing Google Classic library and there are like hundreds of examples from applications or algorithms and you can copy paste the code into Wolfram Mathematica. Try to do the own by yourself. The same example is also in our GUI and we also have the user guide like Docs at Classic IO. You can take the example, you can download them, paste them and just try to copy, paste and do it by yourself. You will see very easy.

And if any question, do not hesitate to ask. You see Ron himself, like, you know, he is very active, you know, asking, responding to any detailed question. And the same like, you know, with Wolfram, like, you know, please, you know, from the documentation, copy and paste the code and ask questions like, you know. Both companies are very committed to help our users with any detailed questions that they may have.

And, and, and you know, talk to us. As we said before, you have the contact details of both companies. We're looking to expand these efforts beyond just work between Classic and Wolfram and we'll be happy to collaborate around numerous interest areas. So please feel free to talk to us and we can get some exciting things on the way.

I'm excited to see what the quantum wish list comes out of this.

Yes, thank you everyone. I'm gonna hand back to Yi. Any final word like you know, Yi from your side.

Thanks everyone for joining us. If you have any additional questions, shoot as an email.

Excellent. And Ron, like, you know, any final word on this, you know, you share your screen like, you know.

Yeah. Yeah, I'm sorry, I just, I thought that just before we finish, maybe the best thing I would allow myself to, to do now and I welcome you to join is I would open a channel in the classic Wolfram Classic community. I welcome you to share whatever you like to ask questions, to put your own examples inside. We will be more than happy.

Excellent. Do you mind like, you know, to share that link so we can also like, you know, pass, you know, put it in the live stream the link on that Slack channel.

Yes.

So doesn't get more practical than this, huh?

We are doing everything like if you put it in a zoom chat, like you know, I'm gonna, you know, share with our team. The, our team, they can take it and put it on. On.

Yeah.

Yes. Okay. Excellent. So we are gonna put it. Okay, so we are gonna put it now on the YouTube channel. Excellent. So please join and ask your questions.

Yeah. Thanks, everyone. We will see you on the Slack Channel.