Liang Jiang (BS '04), Quantum Physicist
January 14, 2022
As an undergraduate at Caltech in the early 2000s, and then returning as a postdoctoral scholar after completing his PhD at Harvard, Liang Jiang was present at the creation of two milestones in the history of quantum information at Caltech - the start of the IQI, the Institute for Quantum Information, and its transformation into the IQIM, the Institute for Quantum Information and Matter. During this period, the field matured from a purely theoretical outgrowth of physics to one sufficiently advanced to incorporate the "matter" provided by experimentalists and engineers.
Jiang pursues research at the cutting edge of quantum information, and he is particularly interested in drawing connections from theory to application in the many facets of quantum science, including sensing, transduction, communication, and computation. His recent appointment as a professor of molecular engineering at the University of Chicago is suggestive of just how broad and interdisciplinary quantum science has become over a time period that matches Jiang's research career.
DAVID ZIERLER: This is David Zierler, Director of the Caltech Heritage Project. It is Friday, January 14, 2022. I am delighted to be here with Professor Liang Jiang. Liang, it's great to be with you. Thank you for joining me today.
LIANG JIANG: Nice meeting with you, David.
ZIERLER: To start, would you tell me your current title and institutional affiliation?
JIANG: I'm currently a Professor at the Pritzker School of Molecular Engineering at the University of Chicago.
ZIERLER: It's an obvious question, but why are you in the School of Molecular Engineering, given your research focus and area of expertise?
JIANG: Well, to be honest, I'm not an expert in molecular engineering in the narrow sense, but I guess the title for molecular engineering is more oriented towards nanoscience, nano-engineering. About two years ago, it was the Molecular Engineering Institute, and now it's the School of Molecular Engineering. Here, we cover both quantum engineering, immuno-engineering, and sustainable engineering. Basically, quantum engineering makes up about a third of the efforts here in the School of Molecular Engineering. Actually, lots of my colleagues are trained as physicists and are exploring quantum science engineering here.
ZIERLER: Is that to say that research in your group is dedicated more towards applications and even translational research?
JIANG: I would say definitely we try to explore novel applications, like overcoming practical challenges to benefit from the power of quantum. On the other hand, as I mentioned, we're also looking at fundamental aspects of quantum information, how to get to the fundamental limits, how much information you can transmit over a quantum channel and how to develop new ways to correct errors. I would say it's a combination of both more traditional, practical science and the fundamental aspects.
ZIERLER: Of course, Chicago has one of the best physics departments in the country. Do you collaborate with anyone in the department of physics?
JIANG: Yes, we actually have a very close collaboration. For instance, I'm working with Professor Dave Schuster's group on using superconducting devices to do quantum information processing.
JIANG: Currently, my group has about 15 people with about half of the members graduate students, half post-docs. Some of the post-docs are post-doc fellows working with other faculty as well. That's roughly the distribution. My group is a theory group working on quantum information. The theme is really to use quantum control and error correction ideas to overcome practical imperfections so that we will be able to achieve the benefit of quantum in terms of quantum computing, communication-sensing, and other applications.
ZIERLER: Beyond Chicago, given the applications that you're working on, where in industry are you doing collaboration?
JIANG: Regarding industrial collaboration, I would say there are probably two types of ongoing collaborations. One is, currently, I'm an Amazon Scholar, so I work closely with researchers at Amazon and AWS to try to explore new ways to do quantum error correction, so we'll be able to achieve scalable quantum computing in the near future. The other potential collaboration, I filed some patents, not necessarily just for quantum computing, but some have ideas for quantum sensing and other quantum applications, and those might benefit other companies interested in using quantum. I would say those are the two general approaches.
ZIERLER: Some overall questions about where you see quantum computers heading. First, just the timeline. There's a range of answers I've heard on this. When do you think quantum computing will be a reality. Are we there now? Are we close to it? Or is it long out into the future, as you see it?
Quantum Computers and the Proof of Principle
JIANG: I think the answer might depend on how you define quantum computing because people have already demonstrated a small-scale quantum computer with a few qubits and even demonstrate some of the algorithms, like a Grover search algorithm of a small scale. In some sense, we've already demonstrated proof of principle, and maybe more of interest to a general audience is how they'd benefit from having a quantum computing device. This is ongoing research, and there are people exploring noisy intermediate-scale quantum applications. Can we use existing noisy devices to do something useful? I think it's happening right now, and it's still an open research question. In the longer term, if we can build scalable fault-tolerant quantum computing, it might take a while, maybe a decade or even longer, but we would be able to build a fully scalable fault-tolerant quantum computer. I think something's happening right now, and it's really exciting to be part of this history.
ZIERLER: In terms of achieving full scalability, as you mentioned, what are both the theoretical challenges, the experimental challenges, and the engineering challenges to get there?
JIANG: I think different researchers have different visions about how to achieve a fully scalable quantum computer. For example, there are different experimental platforms like trapped ions, superconducting qubits, or even optical systems, quantum dots, and so on. I think everyone will have a different version. But I think there's a common challenge, which is that quantum error correction is indeed much harder than classical error correction because quantum states are fragile. We need to, on the theory side, develop efficient, robust quantum error-correcting codes to correct practically relevant errors. This is still an ongoing research topic for my group and many other researchers. Developing good codes and architectural designs will be important.
On the experimental side, I think people are working really hard on improving existing platforms as well as exploring new forms or new qubits. I think scalability is definitely a challenge. You want to build many qubits, which are good, and connectivity is another challenge, you want to make sure these qubits are well-connected in a controllable way. In terms of an engineering challenge, once you have thousands or millions of qubits, how do you control these quantum bits in a hardware-efficient way? And how do you manage potential heating or power consumption? I would say it will probably go to the scale that no single research group at the university level will be able to do it, and it's actually a good time to explore potential industrial or national lab partners to do it at a larger scale.
ZIERLER: You mentioned the idea that quantum systems are fragile. What do you mean by that?
JIANG: The fragility of a quantum system is, first of all, quantum bits are more than classical bits because we want to maintain not only the zero and the one, we also want to maintain the superposition of zero and one. When you have quantum bits sitting in the physical environment, there are all kinds of environmental noise that might cause not only bit flip, but also dephasing, which destroys the interesting superposition that you want to use for information processing, so that is actually really hard. It also makes the error correction hard because we can't look at the bits and do error correction. We need to have clever ways to do error correction without destroying the encoded information.
ZIERLER: With respect to error correction being hard, what are the errors, and what's so difficult about correcting them?
JIANG: Theoretically, we would make a model for the error. Typically, we would say bit flip or dephasing, which is related to whether it's zero or one. Potentially, bit flip would destroy that, and dephasing would destroy the superposition. In practice, there are also leakage errors. It may not just be a two-level system, it can go to some other states. All these are practically relevant errors we need to correct.
ZIERLER: Where you are, coming from an engineering perspective, being in a school of engineering, are you asking different questions? Do you have different interests than if you were in a physics department, for example?
JIANG: Well, I would say, deep in my heart, I'm still a physicist by training. And it's actually an interesting time. Quantum information is being more widely accepted and embraced more by engineering departments and other departments. But from a research perspective, I wouldn't say it's an engineering or a physics question. I think all these questions are important and interesting to address. For the individual researcher, we have our own research tools, research ideas, trying to address and resolve those questions. Of course, there are different types of questions. Some of them are more fundamental, like asking about the fundamental limits, developing general frameworks to address those questions.
Some of them are maybe more specific for a certain type of noise, how to correct it, or maybe for a specific platform, how to design a specific control pulse to achieve the best performance of the gates. I would say, at least for a theory group, we can actually investigate such a wide range of questions, which I personally find exciting, and it's also beneficial to know at the different levels, so we'll have a better understanding and know we're developing a good theoretical model to study the system. Then, maybe we'll have the best theoretical tools to address those questions.
ZIERLER: To come back to the idea of achieving full scalability with quantum computers, to what extent do we need to have a good idea for what quantum computers will be useful for, and to what extent is that not a relevant conversation, and the focus is on, "Let's build them, and at that point, it'll start to become apparent what we can use quantum computers for"?
JIANG: I think that's a process. First of all, we do know there are certain quantum algorithms, like Shor factoring or Grover search, which are things classical devices cannot do. If we can use a device to demonstrate these algorithms at a level that the classical device cannot do, then it's already demonstrated the power of quantum computing. If we can achieve the fault-tolerance and get to that level, that would be great. At the same time, we're also looking into new quantum algorithms. I think that's an important question for the whole community. Because if there is such a powerful device, it would be helpful to further expand the application, not just related to specific algorithms, but maybe we can extend it to some other important questions that people want to solve.
ZIERLER: What does quantum transduction mean?
JIANG: This is actually related to one of the research thrusts in my group. Here, the transduction refers to building a quantum coherent interface so we can convert quantum information from one physical platform to another. The one example is that nowadays, we have quantum computers in superconducting devices, which store and process information in the microwave domain. On the other hand, when we do communication, the best way to communicate over longer distances is to use optical fiber, which store and process information in the optical domain. And the microwave and optical modes are both electromagnetic field, but on the other hand, they work at very different frequencies. It's a nontrivial task to convert microwave photon to an optical photon without losing coherence. This is actually something the community is working on because if we can achieve such quantum transduction, we will be able to connect different superconducting quantum computers without optical fiber link, so we'll be able to build larger, scalable quantum network connecting all these quantum computers to get more computational power.
ZIERLER: Some questions about the idea of quantum communication. What does that mean? Is there such a thing as a quantum internet that quantum computers will change the way we communicate with one another?
JIANG: There is actually a vision that maybe in the near term, we'll build a quantum internet. I think everyone knows we have the classical internet, which sends classical bits over the network. However, that's different from the quantum internet, which we can send quantum bits over. The one key difference, for example, is with the use of quantum internet, we can achieve secure quantum communication, which is a protocol that allows generating secret keys between the users, which they can use for secure communication. In the classical world, there's no such corresponding algorithm to achieve that level of security.
Moreover, with quantum internet, you can also use it for other applications, such as what I mentioned earlier on, connecting quantum computers with the quantum internet to build a larger-scale quantum computer with exponential growth in computational power. For applications, such as quantum sensors, if we can build a network of quantum sensors, we will have a significant benefit. For example, something called a quantum telescope, where you can connect different optical telescopes with the quantum internet to potentially have a better resolution and see more remote stars in the optical domain. There are very interesting applications, which would benefit from a quantum internet.
ZIERLER: The idea of a quantum telescope, to clarify, you're not talking about a telescope that will simulate observation, you're talking about a telescope that will actually see things.
JIANG: Exactly. This is an idea proposed by Daniel Gottesmann and collaborators maybe a decade ago. The idea is that it's difficult to send astronomical optical photons over long distances, but if you have a quantum internet, you can teleport those photons, so that you can effectively increase the size of the telescope, so you'll be able to see more remote objects with better angular resolution.
ZIERLER: That gets me to some other questions about the ways in which quantum computers might help resolve other fundamental questions in physics, like pushing beyond the standard model, finding particles beyond the Higgs, or merging general relativity and quantum mechanics. What are you most excited about in that regard?
JIANG: There is actually a very active research direction right now in which people are trying to use quantum sensors to search for dark matter. Some theories predict axion particles, which might exist, and that building quantum sensor might potentially be useful for that task.
The Caltech and Amazon Partnership
ZIERLER: You're also an Amazon Scholar. To what extent does that get you back to Caltech to collaborate with some people from your earlier years in education?
JIANG: First of all, I think it's a really exciting movement. Caltech is a key partner of AWS research on quantum. Currently, led by Oskar Painter and Fernando Brandão in both experimental and theoretical research. It's great to connect with them. They're working on something big and exciting. Moreover, I find it's really pretty fun working together with John Preskill and Gil Refael, who I've worked closely with during my post-doc, and I've worked with John since undergrad. It's great fun to work and discuss with these people, and it also reminds me of the good old days when I was at Caltech for post-doc and undergrad.
ZIERLER: As part of your research agenda, is the work for Amazon totally separate from what you're doing at Chicago, or do you integrate the research to some extent?
JIANG: This is a tricky question. [Laugh] I think at the fundamental level, we're all thinking about how to use quantum to solve outstanding questions. On the other hand, there is a formality that Amazon research is specific to a particular platform, and they want to do it on a particular device. I would say, I have the other part of the research in my own group, which is more exploratory and maybe more on the fundamental aspects on the different ways of error correction, different applications like sensors and a quantum network. It's interesting to have this balance. It's also really good that some of my students, when they graduate, join AWS as a full-time researcher. It's a very interesting career path for them, working in industry and using their talents to solve outstanding questions.
ZIERLER: Without getting into any sensitive discussion about what you're specifically working on for Amazon, is your sense that these major companies, Amazon, Google, IBM, Honeywell, are involved in a race where everyone is trying to get to the same objective? Or are these different companies pursuing different projects and perhaps even asking different questions?
JIANG: The goals are probably similar, trying to build a large-scale quantum device that can process quantum information. However, different companies are taking very different approaches. For example, Honeywell is trying to do trapped ion quantum computing, while Google is doing superconducting qubit-based quantum computing, while Amazon has a different approach. At Microsoft, there's also a different approach. I think it's actually a very interesting time. Different places are exploring very different methods. Even though the platforms are different, there may be some commonalities in terms of the ideas, how to overcome errors and so on. If there is a breakthrough in one platform, it might potentially benefit other platform and other companies as well. I think that's one good thing, many of these places publish research papers, rather than keeping industrial secrets. I think for this deep tech, it's important that at this early stage, we all work together to overcome these major challenges so we'll be able to achieve the result and the benefit of a quantum system in the near term, rather than working independently and secretly. Because I don't think we're anywhere close to that stage. It's still a very active research area. At the same time, it's really beneficial to have these industrial inputs to get more resources so the field can develop faster and attract more talents to work together on these challenging questions.
ZIERLER: As you know, 50 or 60 years ago, Bell Labs supported fundamental research, whether or not it had anything to do with the corporate bottom line, what was in the financial interest of the industry. Given the fact that today, hundreds of millions of dollars are being poured into this research, does that suggest a similarity, where these companies are supporting basic science? Or are we already at the point where a company like Amazon has articulated ways in which quantum computing might be beneficial to its business, so its supporting research on that basis?
JIANG: I think that's a good question. It may be beyond my pay grade to talk about these longer-term plans for these companies, but I do think it's a good thing that different places not only say, "We're focusing on this particular technique," but when you look at the research publications, they're very interested in these fundamental aspects of research, trying to see if there's a new way of getting a topological order for quantum systems, or if there's a better design for qubits, or if a cosmic ray would potentially destroy a quantum computer, which has sparked a lot of discussion recently. I think these are very interesting things that emerge come this industry-supported research, but I think it's really beneficial for the whole research community to have these exciting results to think about.
ZIERLER: For you, this is unique because most of the people I'm talking to were either only graduate students or post-docs, but you were also an undergraduate at Caltech. To set the stage, when you got to Caltech, that would've been in the year 2000?
JIANG: Actually, I did my undergrad partly in Peking University, then I transferred to Caltech in 2001. Then, I spent three years at Caltech to learn physics, and it was a really fun three years in retrospect. [Laugh]
ZIERLER: At the beginning of your undergraduate education, this is really the beginning of IQI, the beginning of John Preskill starting to think about quantum information. Is that one of the reasons why you wanted to come to Caltech? Was that on your radar, coming from China?
JIANG: When I was applying, I actually wasn't aware that quantum information was an emerging field. I was just starting to take quantum mechanics at that time. However, before I came, I heard there was quantum information research, and it sounded interesting. I was starting to learn about these topics. One thing I'm really grateful for is that when I was an undergrad at Caltech, there was this program called SURF, the Summer Undergraduate Research Program. During the first summer, I worked with Hideo Mabuchi on some of the quantum optics research, which actually was really helpful for me to learn about quantum systems and decoherence. In the second summer, I worked with John Preskill on some of the quantum information questions. I found it to be really helpful. In retrospect, during these undergrad research sessions, I got the feeling they were very different from taking classes because when you take a class, you do homework, you know there is a solution there, you just find a way to solve it. Doing research is the other way around. You don't even know if the question has an answer or if it's even the right question to ask. I found this really, really helpful, having such experiences. Also, I realized I needed to learn more about the field to really contribute to it. I found that a really great experience.
ZIERLER: As an undergraduate, was John Preskill accessible to you? Were you able to talk to and learn from him?
JIANG: Yes, John was very helpful, but still a bit intimidating because I actually didn't know much about the field. But the one thing I recall is that John, in our early meetings, gave me two papers to read and think about. One was about Kitaev's toric code. He, Kitaev, and others wrote that paper together. The other is about teleportation-based error correction, and people in the field know that this is a really important idea. At that time, it took me quite a lot of effort to really digest and understand those papers. But it turned out that during my graduate research, those two papers made a big impact. My research work in graduate school turned out to be based on some of the ideas that John recommended to me when I was still in undergrad. I'm really grateful, [Laugh] and I really admire John's great insight about this groundbreaking work. At that time, it was not so well-known, but it turned out that nowadays, it's a really important tool, not just for my research but for many other researchers' work as well.
ZIERLER: Were you aware of the IQI? Was that accessible to you? Could you hang out there as an undergraduate?
JIANG: Yes, I did. For students, both undergraduate and graduate, free meals are always very attractive. [Laugh] At IQI, there was free dinner at the group meetings. At that time, I was just learning the field, and it was still quite challenging to catch up and know what the individuals were talking about. But it was really good exposure. In retrospect, they're really great, talented people like Michael Neilson, Patrick Hayden, Debbie Leung, and many others. At the time, Caltech was one or two of the centers of quantum information research, in the early 2000s. I was blessed to have the opportunity to get exposed to quantum information at the early stage.
ZIERLER: How much work did you as an undergraduate in condensed-matter physics? Did you appreciate some of the connections that would later be made between the condensed-matter theorists and the quantum-information theorists?
JIANG: I'd say most of the work I do is along the lines of quantum optics, like how to use atoms and photons to process information, which was related to the research efforts in Mabuchi's lab before he moved to Stanford. Later on, the work with John Preskill was more on quantum information as well as how to have ways to build a fault-tolerant quantum computing and achieve robust quantum communication, which later on, my graduate research was also along those lines.
ZIERLER: When you first came to Caltech, did you think you would be going back to China? Or did you intend to pursue graduate school in the United States and make a life for yourself here?
JIANG: That's a good question. I would say when I was coming for undergrad, I hadn't thought that far ahead, where I would be after getting my PhD. But further along in graduate school, I found it was actually very helpful at Caltech to be exposed to the frontiers of research, in particular, quantum information research, which actually changed my entire research career.
ZIERLER: What were your interests by the time you finished undergraduate, and how did that inform what kinds of programs, even people to work with when you started to think about graduate school?
JIANG: At that time, I would say it was still quite open. To be honest, I wasn't even sure if I wanted to do experimental or theoretical physics. But when I got to graduate school, I thought I might even try something else, like bioengineering-related research, but later, I found that quantum information research is still very interesting, and moreover, when I was in graduate school, I worked with Professor Mikhail (Misha) Lukin, and I found there was a combination of both the physical platform and quantum information, and I found that was a good balance point for me with my physics background as well as early exposure to quantum information. I think that's a place where I can contribute to the field.
ZIERLER: When you got to Harvard, did it have a similar center to the IQI?
JIANG: At that time, I would say no because probably only one or two researchers were working on quantum information at Harvard at the time. At the same time, there was the CUA, Center for Ultracold Atoms, with multiple researchers from Harvard and MIT. Over there, they had lots of AMO physicists and some people interested in quantum information who were actually very interesting to speak with. But I would say the biggest focus at that time was on the physical platform. In the group led by Misha, Professor Lukin, there was an emerging interest in doing a quantum internet or even with defect centers to produce a scalable quantum computer at that time.
ZIERLER: What was Misha Lukin working on when you first got to be involved in the research?
JIANG: My first research project with Misha was on the quantum repeater. The idea is that you try to use atomic ensemble to store photons in a quantum-coherent way, so that you can overcome the fiber loss to achieve a long-distance quantum communication.
ZIERLER: What was the process for developing what would become your thesis research?
JIANG: Actually, the theory was developed very quickly. At that time, one thrust of my thesis research was on quantum communication, while at the same time, there was also a development called nitrogen vacancy centers in diamond, which is a color center in diamond that turns out to be a very promising quantum information processing unit, even at room temperature. That's something which is just emerging, and I did some theoretical work based on that, which is actually based on one of the papers John recommended on the teleportation-based schemes. Moreover, even try to sum up the experiments just to demonstrate a simple idea, which turns out to be useful for quantum information processing.
ZIERLER: The title of your dissertation, Towards Scalable Quantum Communication and Computation, in light of what we were saying earlier in our discussion, I wonder if that's a title that could still be written today. Are we still working towards a scalable quantum communication and computation system?
JIANG: It's probably fair to say we're still on the way to scalable computing or communication because now, smaller size systems are indeed being successfully demonstrated. Now, can we go from 50 qubits to 100 or 1,000? Or can we go from quantum communication from 100 kilometers to 1,000, or even global scale? Both are actually in a rapidly developing state.
ZIERLER: Looking back 10, 12, 13 years past, when you started to think about these things to get to that scalable size, what has already been achieved concretely, and what work remains to be done that's as much of a challenge today as it was when you were a graduate student?
JIANG: I would say for computation at that time, it was less than ten qubits, and now, people are planning almost 100 qubits with much better data fidelity, which is important. It's not just the number of qubits, but also the data fidelity. Also, user interfaces are much better now. Amazon Braket is doing that as well as Microsoft. That makes the quantum device more accessible to a much broader audience, including IBM Quantum Experience. That's very different from ten years ago. In terms of the communication, there was the development of a quantum satellite launched by Jian-Wei Pan's group in China, which promised that in the near future, we'd probably have a larger-scale quantum internet over thousands of kilometers. However, this might not be the only way to do it. There are also quantum repeaters. More groups are now developing quantum repeaters with improved quantum memory and more efficient interfaces. Especially in the US, I think it took some efforts to convince the government to get into the field with more theory labs. Here in Chicago, we have Argonne Lab, Fermilab, and University of Chicago. We also have these 100-bit, kilometer-long fiber links, which will allow us to explore quantum internet at the regional scale.
ZIERLER: You mentioned some research that was happening in China. After you defended, were you paying attention to that research in China? Did you think about going back home? More broadly, when I asked you about the race within the United States among these companies, to what extent is there a race at the national level in quantum computing between the United States and China?
JIANG: Regarding that, there was actually a story from when I was doing my PhD. I came up with the idea with Misha to do quantum repeater schemes for long-distance communication. I was actually attending a conference, and I met with Jian-Wei Pan in person. I said, "Oh, I have this interesting result." His group came up with a very similar idea. Later, we both submitted papers, and it turned out to be a good idea that the whole community liked. For that particular example, I think the competition is healthy, at least when people communicate and develop the whole field for these quantum networks. I would say probably the race for quantum technology is global, not just between the US and China. There are also many other players, like Europe, Japan, Canada, Australia, and so on.
I think quantum technology is not just the technology of a particular country, it's the technology of the entire human race. It will benefit everyone. It's different from an arms race. It's not like a weapon, where if you have it, you can attack others. It's more like something you can take advantage to do something better. Of course, I think it's good to have some healthy competition so people try to work harder and boost the improvement of the whole field. But right now, I think it's still a healthy competition between different places. With the industrial companies that have joined, it could become competitive when it gets to the stage where products are being delivered.
ZIERLER: After you finished graduate school, did you consider other post-docs? Or was the opportunity to come back to Caltech too exciting, and it was an easy choice for you?
JIANG: I would say I was fortunate enough to have the Caltech offer, and at the time, I didn't consider many other possibilities. I was kind of aware that I didn't learn enough about quantum information when I was an undergrad, so I needed to come back and learn more about the field to do better research. It was a great opportunity for me to have a second chance to learn more about quantum information and come back to Caltech.
ZIERLER: When you returned to Pasadena, was it already the Institute for Quantum Information and Matter? Or did that transition happen when you were a post-doc?
JIANG: I was a post-doc from 2009 to 2012, and I think the IQIM was approved by the NSF in the summer of 2012. I was about to leave, and if I remember correctly, it had just been approved at the time. But it was really a boost for the quantum research efforts at Caltech.
ZIERLER: So the big transformation really happened at the tail-end of your time as a post-doc.
Putting the Matter in IQIM
JIANG: Yeah, I think so. At that time, I still remember, there were lots of faculty from IQI as well as other faculty, especially in the physics department, having discussions on how to prepare the NSF proposals. I wasn't part of it, but I know people were working really hard to make the proposals successful.
ZIERLER: Obviously, you're coming back to Caltech at a different stage of your career. In what ways was IQI different? First of all, had it gotten bigger since you were an undergraduate?
JIANG: I don't know exactly the size of IQI, but the largest change was that we moved to a new building. The IQI was in Jorgensen, then everyone moved to Annenberg, which was actually a really nice building that I think was just built around 2009, and almost everyone moved over there.
ZIERLER: What were some of the big questions? You said there was more for you to learn when you came back to Caltech. What questions did you have? What were your goals to accomplish as a post-doc?
JIANG: At that time, I was thinking about quantum error corrections and topological quantum systems. Quantum error correction is something I've been continuously exploring, and John is definitely an expert on it. We also discussed a lot and I think took a lot from each other. In terms of topological quantum systems, at that time, it was a pretty emergent topic. People especially appreciated that there was the possibility of using topological systems for quantum-information processing. At that time, I discussed it with John, and we had a project on how to do hybrid topological system that can work together with more traditional quantum bits, qubits, and that was something I found helpful to have John's advice on. The other direction, which related to the topological system, there was an active search for Majorana fermions, which is still an active research topic. At that time, I was lucky enough to work together with Gil Refael and Jason Alicea on this topic, trying to propose some theories on how to detect these Majorana fermions or how to build an atomic system that will support Majorana fermions and potentially could be useful for proof of principle demonstration and quantum-information processing.
ZIERLER: Comparing the early years when you were an undergraduate, just from a palpable level, when you returned as a post-doc, what aspects of quantum computation seemed more realizable, that the progress that had been made was actually getting us to some future where we can envision scalability?
JIANG: There was rapid development during the early 2000s, first in trapped ions, with people demonstrating higher-fidelity gates and so on, and with the design of something called the Sorensen-Molmer gates, which is a robust quantum gate of the ion. For superconducting systems, there was this new design called transmon qubit, which was actually invented by my collaborators at Yale, like Rob Schoelkopf, Michel Devoret, and Steve Girvin. And that allows people to store information in superconducting systems for a much longer time with orders of magnitude improvement in coherence. Moreover, close to the end of my post-doc, there were developments in terms of quantum-limited readout, so that you can read out the superconducting qubits with high fidelity, and that's actually crucial because otherwise, those experiments have to be averaged over many, many times -- which fundamentally limits the power of computing. With those developments around 2010, our experimental colleagues had made significant breakthroughs so that trapped ions and superconducting qubits kind of emerged as the two leading platforms for the field.
ZIERLER: We talked about some of the questions you wanted to pursue as a post-doc. What about what you saw as your areas of strength or expertise? Coming from your graduate program, what were the things you were knowledgeable about, where you felt like you had something to offer in these discussions or collaborations?
JIANG: I think it's actually a very interesting question because I would say every researcher is different and will have their own unique knowledge, background, and skillset. I think for me, at that time, what I brought to IQI was a background of research experiment in AMO physics, atomic, molecular, optics, in quantum optics as well as some research in defect centers in diamonds. Those are probably more on the physics side, so I knew what the realistic noises are, how to model the systems, what controls we could have, the realistic parameters for those systems. I think that's actually helpful, maybe complementary to the things that I learned earlier on the quantum information side. It gave me the opportunity to use these quantum information tools to further explore realistic systems. That's what I saw at the time. I find it's helpful to be exposed to different sets of expertise so that one can actually do new research by combining them.
ZIERLER: Even before the development of IQIM, was your sense that there was already a trend during your post-doc years where the condensed-matter theorists were becoming more involved in the IQI? Was that trend already happening at that point?
JIANG: I think so, yeah. And at that time in particular, Gil Refael and Jason actually came to attend IQI meetings. I was quite impressed when Gil and I talked about their idea of how to look for or create Majorana fermions. Fortunately, I said, "I think I might be able to understand and contribute to that problem." But it was already coming together even before IQIM started.
ZIERLER: What was the research culture like at IQI? Would there be weekly group meetings where everybody would share their ideas? Was there something more formal? Would you do that on an ad hoc basis? What was the community like in terms of sharing ideas with one another?
JIANG: There's a weekly group meeting, which John chairs. There are two parts. First, every group member will give a one-minute talk about their research, any new results or progress on research or open questions, and people might offer some suggestions. The second part will be more of a technical seminar. One of the group members will present results. Later, John imposed the rule that you could only talk for one minute because some group members would get really excited with the results and would talk a lot about their research. Of course, it's by no means a bad thing, it's just for the sake of time. It's a very active environment for people to exchange ideas.
ZIERLER: We talked about competition between companies, even competition between countries. What about among the post-docs? Was there a sense of competition among post-docs? Or was it more collaborative or cooperative?
JIANG: First of all, people have different skills. People often work on different problems, so it is not a direct competition. Moreover, there are collaborations with people working on different topics or who have different expertise. People worked together not just for research, but for job search as well. For example, I shared an office with Alexey Gorshkov, and we were at a similar stage of looking for jobs. We shared information about different places, and some places, we'd both get job interviews and talk about them. Note that this is not a zero-sum game. Overall, if the field does well, there will be more opportunities for the entire community. It's important to keep that mindset so we all work together, make our research stronger to increase the chance of getting good jobs. In that sense, with the mindset that it's not a zero-sum game, it's very important and helpful.
ZIERLER: When you joined the faculty at the Yale Department of Applied Physics, were you joining a large group that was pretty well-established at that point, or was your hire part of a broader effort to build something up from scratch?
JIANG: Maybe a bit of both. Even when I was applying for jobs, in my research statement, I didn't specifically say, "What I will do for the ongoing research over there", it was more on my own research profile about network, quantum control, and so on. On the other hand, when I got there, I did feel like it was such a great opportunity to explore that superconducting platform, although I didn't do much when I was doing my PhD or post-doc, by that time, it was so exciting and a great opportunity. Basically, my research was kind of half developing an error correction scheme for their superconducting platform, in particular, superconducting cavity modes, and on the other hand, I had my own research on quantum network and quantum sensing efforts.
ZIERLER: What did you bring with you to Yale from Caltech in terms of research interests that were really helpful in setting up a research agenda, being a member of a faculty where you have responsibilities to produce things?
JIANG: At Caltech, it was definitely error correction, which I learned a lot from John about. And that definitely helped shape the efforts at Yale, even though they were already thinking about doing it before I joined and had papers on it. But I think some of my expertise also helped in further development. Also, the subject of one of the papers John provided me about teleportation-based gates and error correction. I did some research on distributed quantum computing during my PhD, and that idea turned out to be also very relevant to the platform at Yale University. That was what I found to be very relevant for the research efforts there, which I have a connection to.
ZIERLER: You mentioned some senior people at Yale, Michel Devoret, Steve Girvin, Rob Schoelkopf. What were some of the exciting things they were doing that were relevant to what you wanted to get started on?
JIANG: These are definitely world-leaders on superconducting platforms in both theory and experiment. They did groundbreaking work in finding good qubits, good control, and good readout schemes. At the time, a question would be how to scale a system up and how to do error correction schemes. Regarding scaling the system up, one of the ideas that emerged from our discussions was, maybe we could have distributed platforms and use entanglement to connect different superconducting devices, which I think now, even big companies like IBM and Google are thinking about. That was something interesting I think I made a bit of contribution toward. The other was about error correction. When I joined Yale, they already had some ideas for getting the bosonic cat code for bosonic systems. With some of the early trainings on error correction, I worked together with people at Yale, and we developed further generalized bosonic codes and tried to further optimize the coding, making more work along that direction, including the experimental demonstration of some of the features of these codes.
ZIERLER: I wonder how you might compare Yale's Quantum Institute with the IQI at Caltech. In what ways was it similar, in what ways was it different?
JIANG: I would say it varies in style. I think at Caltech, at least when I was there as a post-doc, it was mostly IQI, which focused more on the theory. At the time, it was mostly a theory center before condensed-matter experts and experimental efforts joined. But I know at Caltech, there are also very strong efforts on the experimental side, led by Oskar Painter, Jeff Kimble, and others. But when I was there, I was mostly focused on the theory aspects. When I joined Yale, they were leading experimental efforts on the platform at the time. That kind of pulled me a bit towards the physical implementation side. That's one place I saw a slight difference. But in terms of the theory aspects, before I joined, it was mostly on the physics side and not so much on the quantum information side. Maybe I brought some expertise on the quantum information, which I learned from Caltech.
ZIERLER: In the way that there are certain corporations and universities that partner up on quantum computation, like Amazon and Caltech, were there particular corporations you worked with or that Yale was partnered with during your time on the faculty there?
JIANG: Yes, actually, my colleague, Rob Schoelkopf started a company with a few people at Yale called Quantum Circuits. They had a plan of bidding some devices for quantum computing. I was not directly involved in that company's efforts, but some of our inventions with filed patents were licensed through that company. Indirectly, I was involved. At the same time, some brilliant students and post-docs from that group I collaborated with actually moved onto that company, so I know quite a few people over there. It's a very interesting effort I guess the difference is that it's a smaller company with a more specific goal rather than an Amazon, which has more resources compared to a smaller company. But the industrial efforts at Quantum Circuits are different than at Amazon.
ZIERLER: What were some of the major papers or discoveries you were involved with during your time at Yale?
JIANG: There were several efforts. One was the one I mentioned about bosonic error correction. We made some theory proposals of how to do universal gates for the bosonic codes and explore that, comparing different bosonic codes to determine which are better as well as developing new bosonic codes. The other is related to the experimental efforts demonstrating something called the break-even performance of bosonic error correction. The idea is, with a real experimental device with realistic noise, Rob Schoelkopf and his group, with some theoretical support, demonstrated that a scheme with error correction can do at least as good as the scheme with the best code without error correction. It might sound trivial, you say, "Well, you just break even," but it actually took decades to get there.
Even now, so far, the only experiment has demonstrated break-even with realistic practical development noise. But I'm optimistic that maybe in the next few years, more platforms will reach break-even or even better. I'm hopeful. Another effort, my research on quantum network, classifies different quantum repeater protocols. The other is about quantum sensing. My graduate student, Sisi Zhou, now a post-doc at Caltech, made a really great discovery, solved a two-decade-old open problem, which was actually a question raised by John and others early on about how error correction could help with quantum sensing. There's also a story about that, if you think that's relevant.
ZIERLER: Tell me!
JIANG: I was attending a research meeting, and I ran into John. At that time, it was 2014, about the time LIGO detected the gravitational wave. Also, there was break-even experimental demonstration with bosonic codes. The natural question was, can we put these together with error correction to help with LIGO? There were clever ways to do it, and I passed the question to my students who had just joined the research group. She was really good. Initially, I thought there was some way we could do it, but she proved me wrong. Later, she even came up with a proof saying it was a no-go, that unfortunately, error correction couldn't help with LIGO against loss error. Then, we further generalized the results with John and other members in the group, and we had a paper about the necessary sufficient condition for error correction to help sensing. That was a really important result and solved a question that was open almost two decades. Of course, the question traced back to John. He really helped a lot with the discussions. I found it very helpful.
ZIERLER: We know what teleportation means from science fiction. What does deterministic teleportation mean in quantum information?
JIANG: The idea of teleportation in the quantum setting is that you can teleport an unknown quantum state from one place to another. There's no violation of no-cloning, because once you teleport, the original state disappears. The good thing is, the state appears in another location, which will be useful for communication or even eventually probably computation. It will be helpful if it's deterministic so that you don't worry about whether something you teleport will be received or not.
ZIERLER: Tell me about your visit to JILA. How was that useful to you? What were some of the exciting things that were happening there?
JIANG: The visit to JILA was actually a combined mission. At that time, I was organizing the Boulder summer school. The Boulder summer school is an annual summer school on condensed matter. The one I'm organizing is the first in that series, first one on quantum information. That particular year, they said, "Oh, quantum information seems to be an interesting topic." We actually had that as a theme for that particular summer. I think it's a great opportunity to invite world-leading researchers. We have 18 of them who come and give lectures. It was also recorded and posted online so more people could benefit from it. That was actually part of the mission, to go to JILA. The other mission was actually to discuss with the researchers there. That's where the JILA thing came from.
ZIERLER: Tell me about your decision to leave Yale and join the faculty at Chicago.
JIANG: It was a tough decision at the time because both places are great for research. At the time, there were multiple factors that influenced the decision. First of all, I really enjoyed collaborating with my colleagues at Yale, and we're still collaborating. Probably, the major influence was actually that my wife got a research assistant professor position at Chicago. For her research, it was a unique opportunity for her where she could start independent research. That was a major factor. Mostly, family reasons. Research-wise, there were a few benefits. There are quantum network efforts here at the DOE labs I mentioned earlier and more platforms related to quantum sensing. There were multiple factors. But I really liked working with people at Yale, and we still have frequent discussions about different research ideas.
ZIERLER: Given how important collaborations are to your research, did moving to Chicago change your research at all, the kinds of things you were interested in and wanted to pursue?
JIANG: I would say it gave me an opportunity to further expand my research. For example, on the quantum sensing efforts and working with the experimental groups here, that probably will be better quantum sensors with some ideas. In terms of a quantum network, here, they're closer to the test bed of quantum networks, and I have students and post-docs working in that direction.
ZIERLER: Bringing our conversation right up to the present, circa January 2022, what are you working on right now?
JIANG: I'm actually trying to finish a few papers. We have research results which show that we have a quantum scheme. We'll file a patent, submit a provisional next Monday. I assume that's OK to say now. [Laugh] We've found a quantum-inspired classical algorithm, which can do something better even than quantum computing. [Laugh] Which is interesting. I don't know whether to be happy or unhappy.
ZIERLER: That's a pretty unexpected result, I'd assume.
JIANG: It's unexpected. Initially, people were saying this would be one of the applications for boson sampling. It's an algorithm that allows you to get to the molecular vibronic spectrum of the molecules. It's pretty difficult with the existing classical algorithm, and people are proposing using a quantum device to do it. My post-doc, Changhun Oh, who pioneered the research, has come up with a great idea inspired by using quantum optics techniques combined with some other methods to allow us to solve the problem with comparable or even better performance than a quantum device. I have mixed feelings because I really want the quantum device to have the advantage, but at the same time, it's really good to know that there are promising classical algorithms that can do it as well, inspired by quantum.
ZIERLER: Even by way of accident, one of the fun things with fundamental research is, you don't know where it's going to take you. Is one of the promises, then, of quantum research that it might actually make classical computing better?
JIANG: Yes, I think it's possible it could improve classical computing in multiple ways. I know one possible approach is using these optical circuits, you can significantly reduce the power consumption and the heating problem, which is important. I know people are thinking about novel optical ways to reduce power consumption with quantum-inspired technologies. I think that's another way of helping classical computing besides using inspired new algorithms or even a quantum GPU to help with classical computing. Even these basic elements of classical computers could benefit from the quantum technology.
Quantum Computing as a Benefit to Classical Computing
ZIERLER: Looking to the future, as you well know, there's always the element of hype in quantum computing, that it's just around the corner, that it's already here. Where do you see the hype, where we should tamper expectations, and where do you see the excitement that this is real, and we're actually on our way to this goal?
JIANG: It's great that people have enthusiasm and high expectations for the field. On the other hand, I think it's the responsibility of we researchers that when we make claims or predictions, they should be based on what we're thinking. Sometimes, there's a tendency to say, "My research will solve the whole problem," and so on. But we should be cautious in making those claims. Overall, the field has many interesting promises. On the other hand, I think many of them are actually open research questions, and we should be careful in making claims. At the same time, it's a great time to do research, and we're working as hard as we can to solve those problems, and hopefully, we'll be able to have a really large-scale, fault-tolerant quantum computer in the near future.
ZIERLER: For wherever the research takes you, what are some of the things you learned at Caltech that will always stay close to you, that will inform your approach to the research?
JIANG: There are multiple things on different levels. In terms of trying to solve different levels of research problems, many of my mentors at Caltech really reach out to solve important problems, which I find very helpful. All those discussions are very encouraging and instructive. I think that's very important for research. The other is, maybe more on a technical level, the skills I learned from John Preskill, Hideo Mabuchi, and other mentors when I was an undergrad, like Bob McEliece, who was a well-known professor at Caltech working on error correction. I was fortunate enough to learn about classical error correction and classical channel theory from him, which I'm really grateful for. Also, David Politzer, who was my academic advisor, and Kip Thorne, who I took multiple classes from and was always very encouraging as a role model. Also, Alexei Kitaev, even though I haven't collaborated with him. He's really an excellent researchers making great contributions to the field. Having the opportunity to just work with these people and learn from them, even daily interactions, is really encouraging and inspires me to do better research.
ZIERLER: Last question. Because the development of quantum computers is more of a process than a singularity, where there's a before and an after, as we continue on this process toward quantum computation, what are the benchmarks you'll be looking for, both as a participant and as a watcher among your colleagues in this pursuit?
JIANG: For different thrusts, there are different milestones. For example, for the quantum sensor, I think we're actually already having lots of interesting quantum sensors that allow us to probe nanoscale spin dynamics with high resolution. That's something I would say is already emerging as a promising application, which might already have some industry applications. In terms of communication, I think longer-distance communication because you can already do communication over 100 kilometers. If we can push it to 1,000 or global-scale with high communication rates, that will allow us to have a broader use of the quantum technology, which might affect everyone's daily lives. That also opens the possibility of building a larger-scale quantum network. For computing, I would say it's more the size, quality, and controllability of different qubits. I really look forward to seeing how error correction can help improve the performance of quantum devices. Once that's there, then we can start to build more and more so we can have a fault-tolerant quantum device. I think that's an important milestone which will emerge as an outstanding challenge for the whole field in the near future.
ZIERLER: It sounds like no matter what happens, it's going to be exciting along the way.
ZIERLER: Liang, it's been great talking to you. I'm so glad we were able to do this. Thank you so much.
JIANG: Thank you for the questions. It's good to think about all the nice days I was at Caltech! [Laugh]