Podcast with Chris Ferrie, Prolific Author and Professor

My guest today is Chris Ferrie, an associate professor of physics at the University of Technology, Sydney and a prolific author of books teaching scientific concepts to children. Chris and I spoke about whether qubits spin the other way in Australia, his teaching and research interests, the quantum ecosystem in Australia and more

Listen to additional podcasts here

‍

THE FULL TRANSCRIPT IS BELOW

Yuval: Hello, Chris. And thanks for joining me today.

Chris: Thanks for having me.

Yuval: So who are you, and what do you do?

Chris: My name is Chris Ferrie. I'm an associate professor at the Center for Quantum Software and Information within the University of Technology Sydney in Australia. We just know it as UTS. And I do research in some areas of quantum computing. And also, I think I'm probably better known as a children's book author. So, I've written some books, including quantum physics for babies.

Yuval: And you've written dozens of books, right? On many aspects of physics, I think.

Chris: Yeah. In English, I've written, I think we're at 60 books now. And that includes now that my own children are getting older, I'm getting less interested in writing baby books and so I've started writing some adult nonfiction books as well.

Yuval: Wow. And so you've taught babies, and you're teaching graduate students, and college students. How about in between? I mean, do you have experience or interest in teaching high school students or grade school students? When is a good time to start with quantum?

Chris: That's a great question. I mean, there's actually a kind of boring technical barrier, in that. To teach babies is in some sense easy because you just have to connect with parents. And they're the ones that buy books and show up at libraries and events. As a professor, you kind of have the freedom to do whatever you want. So you can teach people that show up at university. But in the formal public education system, there's a curriculum. And if that curriculum doesn't include quantum computing, then it's very difficult to even get your foot in the door. So I would say, yes, we should be teaching high school students, and certainly even elementary school students the basic aspects of this technology that may be important in their lives in the future, or maybe they want to be involved in helping build it. But it's not in the curriculum, and so it becomes very difficult to just get in front of that audience.

Yuval: And what would you teach? Do you want to assume that they know trigonometry? Would you teach them matrix multiplication from the beginning? How would you go about it?

Chris: Well, I mean that, yeah, that's a good question, I think. How would I start? I don't think I would start with mathematics for younger audiences. I mean, it's much the same way that we teach students, young students, facts about the world in the universe that there's really no direct evidence for. Say that everything in the world is made of atoms. Everybody knows that, nobody questions that. But do you know what the evidence for that is? So it's not like we have to teach them quantum field theory before they're convinced that atoms exist. And in much the same way we need to find out what the important things to tell people are about quantum technology, so that it becomes like atoms. That is just something about it, that everyone just agrees is true and they can get sort of get on with the next steps.

So that's, I think, where I would start is to say that there's a different sort of way to do computing. And this will help us solve some new kinds of problems, like in designing materials the kinds that we enjoy that are completely unnatural that we're surrounded by. I'm surrounded by basically nothing that's natural in this room. And new medicines, which I think include things like vaccines, which I think are entirely relevant today. So I think it's easy to connect aspects of quantum computing to your everyday world. And then I think in high school in much the same way as we start to introduce math into physics, then we would introduce the concepts. But I think I would start with just circuits, just quantum circuits. Just say, here's a thing and you can add gates to it and it does stuff, let's play with it and see what happens.

Yuval: What do you think is the most difficult concept for people to understand? Is it the superposition? Is it entanglement? What is it about quantum that becomes spooky?

Chris: So I think the meta difficulty is it is actually for experts to accept that we don't need to teach people about superposition and entanglement, in much the same way that we don't need to teach people about tunneling in transistors before they understand how to code a web app. Right? So when people interact with quantum technology in the future, they won't know what superposition and entanglement are. They'll have some higher level language that connects with the standard interfaces that we have; keyboards, and monitors, and mice. And there'll be some abstractions which don't require that people understand these sort of nitty gritty details about physics.

Yuval: Excellent. Education aside, you are also a researcher. Could you tell me a little bit about what you're researching? What are you trying to find? What your students, more or less, are working on?

Chris: Sure. Yeah. So I am a researcher. I have a large research group, which means I'm more of a manager. Unfortunately, I don't get a lot of time to do my own research, that my own personal interests are, and follow what the grant milestones tell you that you need to produce, which is a few things. Maybe I can tell you in the context of the quantum software stack. Right? So we work at sort of the low level hardware instructions, also known as quantum controls. Trying to design the really sort of low level instructions that you would send to some piece of quantum technology, a quantum device, whether it's natural, like an atom or artificial like a superconducting circuit, to get it to do what you want it to do.

So part of that, so there's a sort of a catch-22 there that if you are just presented with a device, you don't know what a description of it. So the first thing you need to do is characterize it. And the way you do this is, again, through sending the instructions and then looking at the response. So we do both the characterization, and control of quantum devices. Not real devices, like at abstract level and using the theory of quantum physics. Some of our protocols and results end up being used in laboratory, but we try to make them sort of general. And then moving up the stack, we work on hybrid quantum algorithms. So maybe the listeners are familiar with quantum optimization, things like QAOA and VQE and lots of other sort of three letter acronyms.

And what we try to do there is see what the effect of noise is when you try to do a very detailed simulation of these algorithms. So for a real problem, say, take a vehicle routing problem or a traveling salesman problem, or something like that, and break it down into all the gates that you need to run on a quantum computer, but then include noise and see what happens. So it's very sort of in the weeds and practical in that sense. And then at the sort of top of the stack, we work on quantum machine learning, and in particular quantum algorithms to train quantum neural networks and other similar sort of primitives. So rather than what people envision today is using classical computers to train quantum neural networks. And that has a lot of problems associated with it. So we are designing quantum algorithms that will train these networks. And of course that requires a large scale quantum computer. So this is sort of more blue sky stuff. So we kind of cover a lot of the software stack.

Yuval: When you talk about pulse shaping and the low level stuff, is that a transient problem? Meaning, do you expect that three years from now, people will still be looking at pulse shapes and how to get atoms to do what you want from them?

Chris: I think so. I think a lot of it will move into academic research and labs. People that have prototype devices building new technology for quantum computing will have to continually innovate and design new control systems for them. So I think that will always happen. My suspicion is that in industry, you probably don't want a system that requires a PhD or a whole team of PhD, quantum physicists constantly by hand designing pulses. So there, I think there are be it some stable system or some actively stabilized system that works in an automated way.

Yuval: One of the things that we think about at classic is scaling up systems from five qubits to 50 qubits, to 500 qubits and not just pulses, but the algorithms. How do you create an algorithm that is efficient and makes efficient use of 500 qubits or whatever large number when it's becoming difficult to do so by hand. What do you see as the issues of scaling up? I mean, toy circuits are nice, but ultimately, we are going to need bigger computers and bigger circuits to solve big problems. What do you see as the main issues in getting there?

Chris: Well, the problem of course is the exponential scaling in quantum technology. Right? So in the sort of typical model that we think of, we assume we can control individual qubits, which we need to do, and then we can control a few interactions between qubits. And then we have to build algorithms on top of that sort of restricted connectivity or topology. And in some sense, it kind of has to be that way because you can't design a control system that will create a single interaction amongst 500 qubits, not in the way that we do it now, which is we design them with conventional computers. So with 500 qubits, you just can't simulate it. Right. So I think that's the problem with control is that we're in this classical mindset. Right? We have to... We feel the need to be able to see the solution. And that means simulate the model, and that you just can't do once you reach a certain qubit number.

Yuval: How much are you involved, or your colleagues are involved with industry? I mean, do you get 17 calls a day, "Hey, help us shape this pulse." Or, "We want to build this new quantum computer, or do better VQE." Or are you just happy working in the research lab?

Chris: I mean, it's a bit difficult in some Australian... The problem in Australia is that every university has its own IP policy for its staff. And so it becomes very difficult to engage between industry and academia. Especially nowadays when universities in Australia are struggling because of the pandemic and the lack of international students, and that means lack of revenue. So, unfortunately they're looking to say protect more aggressively protect their IP. And so for a lot of researchers, it just becomes a nonstarter. They just don't have the time or patience or energy to deal with their own university's lawyers and whoever else sits on the top floor.

So, yeah, there's not a lot of interaction. Most of the interaction I see is between former colleagues. Right? So we have many colleagues that have decided that they're going to leave academia, maybe even only temporarily to go into industry and they maintain what was their former academic connections. And so, yeah, in those cases, there's a lot of informal conversations that happen, but not a lot of formal stuff. There's not a lot of contract research that's being done in Australia between industry in academia.

Yuval: So I think you and I were both wondering if qubits spin the other way in Australia. But other than that, from what you can see, how is the Australian ecosystem different than other parts of the world? Whether it's the way the government interacts with companies, whether the balance between companies and academia, or is there anything that you could share with us that you think is unique about the Australian quantum ecosystem?

Chris: So it's been tough I think with the pandemic, because Australia is just so isolated. I mean, physically, obviously, it's isolated. But that means that even though we have the internet and... I mean, we're evidencing it right now that the world is quite small. But with these conversations don't happen a lot. Right? You tend to interact with people that are physically close to you most. Right? And so I think in Australia, we've kind of lost a bit of touch with the ecosystem globally, for that reason just because we tend to interact with the people that are local and we're not traveling to make ourselves non-localized.

So the Australian ecosystem, I think, is much smaller from what I can see from all the way across the ocean. And it's certainly not growing at the same pace. And part of the reason I think is because, around the world many governments have national or multinational initiatives around quantum technology, which includes a large amount of funding. In Australia, we don't have that. There's no global or no national quantum initiative. Although there's always sort of talks about getting one started. So I think... I mean, that's not surprising if you're in Australia, it tends to sort of lag behind a little of it, which isn't bad. We get to see what works and what doesn't work, and then cherry pick the things that do work. But it's, yeah, as I said, the there's only a handful of start-ups in Australia and of course, multinational companies have a little bit of a footprint in Australia, but not to the extent in the rest of the world.

Yuval: As we get closer to the end of our conversation, let's assume that you are master of the quantum universe for a couple years, and you control the activity of companies like Classiq and other commercial companies in this space, hardware, software, and so on. What would you like us to be working on in these two years that you're master of the quantum universe?

Chris: Yeah, I want you to build a quantum computer if you don't mind. I think, I mean, my point of view is that there's seems to be a problem in that the ecosystem or the people trying to enter the ecosystem with money, and a lot of the people within the ecosystem with egos seem to think that quantum technology, including the task of building a quantum computer, is like other Silicon Valley technologies that... So somebody in their garage can just come up with a good idea, and it'll be some unicorn company that'll make a handful of people really rich. But I think the reality is that the task of building a quantum computer is probably the most difficult scientific and engineering challenge that humanity has ever faced. It's more difficult than landing on the moon. Right? It'll be more difficult than, well, we've already landed on Mars, but landing a human on Mars, I think will be easier than building a quantum computer.

So I think we need a more coordinated effort, not a bunch of... Well, what it was 10, 15 years ago was a bunch of individual academic groups, maybe a handful of people where most of the work was done by graduate students that came in not knowing much. And now it's just changed to be done by a handful of companies with a little bit more money and a few more people. That's not how it's going to happen. I mean, it will take a coordinated effort amongst a large number of people with a large amount of funding. And so if I was the master of the quantum universe, then I would just bring them all together, and yeah, try to make it happen.

Yuval: Fantastic. So, Chris, how can people get in touch with you to learn more about your work?

Chris: I think probably the best way would be Twitter, I suppose. So it's @csferrie. Or, yeah, I guess you can check out my website, or Google me, and there's some contact info there.

Yuval: Excellent. Well, thank you so much for joining me today.

Chris: Thanks for the chat.