Interview by PillarQ&A: Alexandra Boltasseva on Research and Workforce Development at the Intersection of Quantum and Artificial Intelligence

PillarQ&A: Alexandra Boltasseva on Research and Workforce Development at the Intersection of Quantum and Artificial Intelligence

By Katie Elyce Jones, PillarQ

Alexandra Boltasseva is the Ron and Dotty Garvin Tonjes Professor of Electrical and Computer Engineering at Purdue University in the Elmore Family School of Electrical and Computing Engineering (ECE). Boltasseva works at the intersection of engineering, quantum science and technology, and artificial intelligence (AI). She is also the workforce lead for the U.S. Department of Energy’s Quantum Science Center, a national quantum information science (QIS) center in which Purdue is a key partner.

Q&A with Alexandra Boltasseva

Editor’s note: This Q&A has been edited and condensed.

Could you tell me a little bit about your professional background and current research?

My PhD is in electrical engineering but my specialty during my PhD was working with so-called nanophotonics. This is an exciting and contemporary part of optical technologies where we gain control of light by building nanoscale devices that enable novel technologies. That could mean disruptive, new applications but also existing optical technologies scaled down to the nanometer scale so that we can make devices more efficient and compact and integrate more functionalities onto a single chip. These applications could ultimately enable faster and more powerful computers, biomedical sensors, new imaging techniques, nanofabrication techniques, energy applications, and emerging quantum applications.

At Purdue, my collaborators and I started to explore the overlap between nanophotonics and quantum technologies quite a few years ago. We have a big nanophotonics team and a big, center-level activity called the Purdue Quantum Science and Engineering Institute led by Yong Chen. My part relates to utilizing nanoscale structures and modeling properties of light relevant to quantum science and technology, with the aim of enabling quantum devices such as single-photon sources or sources of entangled photons that would be a critical component for quantum communication systems, quantum sensors, and future quantum photonic computing.

You’re also a principal investigator of an Emerging Frontiers Center in quantum and AI. How do the institute and the center relate to each other?

A big theme of our engineering department is building a coherent effort between what we call virtual and physical parts. The physical part means the size and technology as related to hardware devices, and the virtual part means the algorithms and software—the computation for building novel technologies. Our Elmore ECE Emerging Frontiers Center Crossroads of Quantum and AI was built on the idea of merging two technology revolutions that we’re witnessing right now: the revolution in quantum science and technology and the second one, which has been ongoing for some time, is the big data and artificial intelligence or machine learning [revolution]. This is a great example of building a bridge between the physical and virtual parts. So, the Elmore Emerging Frontiers Center for Quantum and AI has a very nice overlap with the Purdue Quantum Science and Engineering Institute.

In one of your biographies online, you mention that “Many future technologies are not single discipline.” Why are we now seeing the merging of disciplines like electrical engineering, AI, and quantum?

Every time we think about what’s next, we analyze what led us to this next step in technology and science. If we look at the first Information Revolution with the invention of the laser and the rise of the Internet, now we are witnessing the second one, which is [defined] by big data, AI, and virtual reality everywhere. At the same time, we are witnessing advances in fundamental science and quantum technology. For example, the Nobel Prize this year was given for quantum science and technology. Experiments of non-classical effects and advances in quantum materials and, to that extent, nanotechnology, have led us to the point where we can say, ‘Ok, now we can really push it.’”

It’s interesting because this revolution is going to be bigger than the revolution [enabled] by the introduction of transistors and lasers because we are leveraging what we already have. The previous revolution happened as fundamental discovery—first at the university and then it slowly propagated to the semiconductor industry as we know it today. But for the quantum revolution, everything is happening concurrently. It would be fair to say that a large body of fundamental research is happening at big companies like Google, Microsoft, and IBM that are building quantum computers and quantum sensors. The fact that this revolution is happening concurrently at universities, national labs, and industries is very important because it allows us to push the frontiers of both science and technology really fast. An example is the DOE centers [the National QIS Research Centers].

What are a couple of critical ways these disciplines are merging?

First, if you did a Google search of “quantum” and “machine learning,” most likely you will get something related to quantum machine learning. This is a new direction where you utilize quantum solvers or future quantum computers for artificial intelligence. Right now, it is a very big field. People are working hard both on building the machines but also on thinking of the algorithms they will use on quantum computers. Quantum computing promises to solve very complex problems faster than classical computers. How this would advance the area of AI is very interesting. So, developing algorithms for quantum machines to be able to solve what’s now being solved on classical machines is one important area.

Another important area is what the Elmore Emerging Frontiers Center is pursuing—which is how we can use classical machine learning algorithms to advance the area of quantum science and technology now and not wait until the quantum computers are around. We can do this in multiple ways. First, designing [quantum] devices is one example. We’re using machine learning algorithms in a similar way that is used for solving the inverse design problem. Inverse design is when you have a function that you want to get from your device, but you don’t know what kind of design will give it to you. [Machine learning] algorithms help us calculate this much faster.

Discovering new quantum materials is another. Performing quantum measurements is a very sophisticated and time-consuming technique, so it is critical that we learn how to train our machine learning algorithm like a classifier to quickly scan our sample and see with high fidelity what is going on. To come up with new ways of studying materials, you need techniques that are very efficient, and algorithms would be not just instrumental but disruptive. Similar to the revolution that is happening with image recognition, we’re trying to adapt this and use it in quantum science, which has never been done before.

For students who are interested in working in emerging areas like quantum, are they still getting traditional science degrees? What kind of choices are they making as they go through school?

This is a very important and not easy-to-answer question. In fact, the Quantum Economic Development Consortium started to look into workforce for quantum and what kind of skills are needed to contribute to the quantum revolution some years ago. This is not an easy task because it is a multidisciplinary field. Right now, there are many different possibilities to work in quantum with different qualifications. Having a PhD in quantum physics is an obvious one, but there is really a need for engineers who are trained both in the physical part—which means semiconductor devices, nanomaterials, nanofabrication, electromagnetics, low-temperature physics—who also have understanding, at least at the basic level, of algorithms and computer engineering disciplines.

Even if they are pursuing a conventional degree like quantum physics, we recommend our students take classes in machine learning or algorithms because there is no way in the future not to have this overlap. Quantum curriculum is something that many universities have an interest in and are building efforts around. For example, the National Science Foundation, the Department of Defense (DoD), they’re investing in new centers. At Purdue, we have a new DoD-funded project in quantum pedagogy, or education, which is focusing on quantum curriculum development.

Are there any other challenges to working on research at the intersection of these fields? Final thoughts?

In addition to the workforce challenge and training, I would say there is another opportunity. The opportunity is that all of us have to work on building what we call the quantum culture. Semiconductor devices are now so common that no one really thinks much about it. We can relate to the technology. There should be some kind of—I don’t want to call it demystification of quantum—but something that people are excited about but not scared of. For many people, the quantum field sounds so far off that it will never happen. We need education of the general public. We need kids—we need girls—to go into quantum.

We’re going to have both quantum and AI be a part of our lives. I think it’s a little easy with AI because it’s in games and smart homes. Having quantum entering our lives to that degree is exciting but it will take some time. We do already have quantum technologies around us, for those who say there are absolutely no quantum applications yet. We have quantum sensors, quantum imaging—they’ll be around big time. And secure quantum communications will be the next step, which is something everyone can relate to because we’re all connected and use the Internet, so our ability to make it safe is something understandable for people.

But let’s leave some mystery for quantum still because it’s an exciting and a very rich field.