QED-C is leading U.S. quantum technology to industrialization

Building the Quantum Information Science (QIS) industry is Celia Merzbacher's job as Executive Director of the Quantum Economy Development Consortium (QED-C). QED-C was established under the National Quantum Initiative Act of 2018 and is broadly overseen by NIST and the National Office of Quantum Coordination.

Merzbacher is a materials and nanotechnology scientist who brings together research, business and government experience. "It's the perfect job for me," she said.

At the White House Office of Science and Technology, she oversaw the National Nanotechnology Initiative; she also served as Executive Director of the President's Council of Advisors on Science and Technology (PCAST); at Oak Ridge National Laboratory, she was Director of Strategic and Institutional Planning; at Semiconductor Research Corporation, she was Vice President for Innovation Partnerships; Merzbacher was elected a AAAS (American Association for the Advancement of Science) Fellow in 2021.

Celia Merzbacher

Promoting the development of quantum computing is only part of QED-C's mission. Understanding the technical and commercial needs of the entire quantum field and accelerating its progress are their main goals. Recently, Merzbacher spoke with HPCwire about related challenges and opportunities. Below is part of the interview, covering technical bottlenecks, workforce issues, governance and coordination challenges, and the immediate priorities of QED-C.

HPCwire: Let's start with the background of QED-C.

Merzbacher: First, we're not just computing. What we're really looking at is all the applications of the second quantum revolution. The National Quantum Initiative Act of 2018 requires NIST to create a stakeholder consortium—QED-C. We got start-up funding from NIST, but we're really industry driven. Today, we have about 200 non-government members, mostly businesses, so about 150+ businesses of all sizes, from all parts of the quantum ecosystem, or supply chain. This includes measurement capabilities, equipment, lasers or electronics and other non-quantum but quantum-enabled secondary and tertiary suppliers,

You can see all our current members on our website: IBM, Google, Microsoft, Amazon, etc. Of course, these companies are not pure quantum companies, they all have quantum departments within them. So all of our members are kind of like a small company in a sense. We also have consulting firms like Deloitte and Booz Allen.

HPCwire: The quantum ecosystem has expanded dramatically in recent years. What is the mission of QED-C? How do you perform this task?

Merzbacher: Our mission is to advance and grow the quantum industry. QED-C is a very bottom-up organization. It's very streamlined and we have a number of committees where members come together to do activities. Include use case committees, understanding the market and size of applications to those markets, the readiness of the technology, and the time it will take. In general, we have a lot of information about where things are and where they are going, but we don't have all the answers yet. Internally, there is a lot of information being shared. I hope that over time we will be able to publish more digested documentation to help the public and community get reliable, credible information that sometimes wonders: is this just hype? I still don't get it. Where is it going?

We have the ability to help educate and explain. Over time, I hope we can do this with the use case work we're doing. For example, a workshop was recently held that will soon report on quantum computing for the grid. This is an interesting use case; let's dig a little deeper and gather in that community, who may not be quantum experts, to see how quantum could work in their field. There's a committee focused on enabling technologies, and we've done deep research and developed some roadmaps in cryogenic space, laser space, electron space, and many different non-quantum technologies that are being used for quantum applications. For example, there are no off-the-shelf lasers or electronics that really meet these specifications today. Many companies either test, modify, or sometimes make their own. Frankly, there's a big supply chain problem and we're trying to identify those gaps in order to fill them.

HPCwire: In terms of your point about supply chain, the overlap with the DOE work is interesting. I recently talked to quantum researchers at Oak Ridge National Laboratory, and one of the things they're doing is trying to develop more efficient sources of single photons that would be useful in quantum networks.

Merzbacher: If you go to our official website and scroll down on the login page, there's a button that says TAC (Technical Advisory Committee)...you can see descriptions of all the different committees.

HPCwire: In the HPC (high performance computing) community, attitudes to quantum technologies vary. A few were excited, most recognized the long-term potential, but some became a little numb to the string of commitments. What are your thoughts on bottlenecks? What do you think is holding back progress?

Merzbacher: As you pointed out in the comments, the various stakeholders in quantum, the bottlenecks are up and down the stack. So if you're talking about the sensing capabilities of the navigation system, it has different technical bottlenecks and requirements than IBM's superconducting quantum computer. And IBM's superconducting quantum computer has a different bottleneck than Quantinuum's trapped ion system. It's really a wide field.

The other thing I would say is that, at the same time, all these activities are taking place in the private sector, as well as in the public sector, government projects and projects that will be invested in. For example, Oak Ridge or NIST. Therefore, these investments require an understanding of the private sector as well as the needs of the industry. This is where the QED-C comes into play by conducting these gap analyses in the workshops where we conduct in-depth research. We have done this on single-photon sources and detectors. Not so much what the technical requirements are, but rather a step before that; it turns out that there isn't even an agreement on how to measure and characterize these light sources. And Oak Ridge will never be a manufacturer of light sources and detectors of any size. They are research institutions, and they try to push the boundaries of capabilities.

So this is a very different requirement than your standard telecommunications system. You need to really have unique ways to characterize, measure, and specify technology, and that's not even agreed upon. So a lot of what we're talking about is the need for a standard, but not a final standard, just a standard language to describe a single photon source or detector. This is a business question, not a real fundamental research question. QED-C is trying to connect the world of basic researchers, and what they're doing, that's pushing the boundaries of science, and those trying to figure out how to translate that into business.

HPCwire: OK. How does QED-C manage itself?

Merzbacher: We have a steering committee of members elected by members. We get a small amount of funding from the government, but we are not a government agency, we are not controlled by the government, we are controlled by our members. I have 200 bosses. The Steering Committee has four representatives of small companies and three representatives of large companies. Small companies are interested in taking more seats, and government agencies have two. Today, the latter two seats happen to come from NIST and the Department of Energy. I would point out that QED-C isn't actually even a statutory thing, it's actually managed by SRI, so I work at SRI International.

HPCwire: I didn't realize SRI was involved.

Merzbacher: Yes, it stands for Membership and Government Management Alliance.

HPCwire: QED-C is relatively young. How do you measure your progress? What are the milestones for this year and next year?

Merzbacher: It's challenging. In a sense, we do what a lot of industry associations do, and I don't know how they measure their progress. We're not a lobbying organization -- that's a bright line because I work for SRI, which does a lot of government contracting work. So we don't lobby, but we can educate. We certainly want to educate government policy makers about what's going on in the industry so that they can make informed decisions. That means we go to educate examiners at patent offices, or meet program managers at the Department of Defense or the Department of Energy or the National Science Foundation, and say, "Hey, there's a lack of basic understanding in this area. In addition to all the other things you do , which is fundamental research that the industry wants the government to cover.”

We also do a lot for our small member companies just to help them understand that as a business they need to know a few things. Some of them are not quantum. For example they need to understand what compliance is important when entering into government contracts, or point them to funding opportunities and help them find interns or summer jobs because we also have university members. We connect university students with companies with jobs. We do a lot of different things to try and fix the bottleneck. We are also working on benchmarks and standards. There are some long-standing benchmarks for high performance computing. How to measure the progress of quantum computing today? We also have people starting to do some work in this area.

HPCwire: What report did QED-C publish? Do you need to submit an annual report or quarterly update (at least to NIST)? Are those public documents?

Merzbacher: We meet with NIST a lot. We have a contract between SRI and NIST, so there are reporting requirements. The documents we provide to NIST are not public, but we do disclose some material, which is typical of a membership-based organization. Members pay to become a member and get some benefits as a result. Some of our reports are only shared among members, some are public like our recent release, which may actually be interesting, and are requirements for an intermediate representation or abstraction layer between hardware and software and quantum computers. We ran a workshop that took a deep dive into this. We decided to make this report public because we wanted the world to think about how to run different software on a lot of different hardware.

Here is an example of a report we published. Last fall, we also published an article called "A Guide for Quantum-Safe Organizations." It's more aimed at what I'd call the long-suffering CIO responsible for the security of corporate IT systems; they've heard of quantum computing, but aren't sure when it's coming and what they need to worry about. We released a report to try to educate those in these roles what the threat of quantum computing is and what they should think about today to prepare. This is a special aspect of quantum computing - it has the potential to crack encryption.

HPCwire: We haven't talked too much about when quantum information science will start delivering concrete benefits. In quantum computing, the race is to achieve quantum supremacy on NISQ (Noisy Intermediate Scale Quantum) computers. What do you think?

Merzebacher: I think quantum advantage is something everyone is excited and eager to see, and it's challenging for a number of reasons. Building a sufficiently powerful quantum computer requires overcoming some very big problems. You have problems with everything: like the quality of the qubits, their fidelity and their connectivity. Then you have problems with error correction, environmental control, and scaling. If I have to say, scaling is the biggest issue for the next year or two. Even if you can show something in the lab and you've got great results, how do you put it into manufacturing and scale it up?

Another problem is that your goals are constantly changing as existing HPC advances. When you're doing this kind of crossover, it's hard to predict the future. People in the lab, in places like Oak Ridge, more professional than me, say quantum computing is going to be like an accelerator, just like a GPU is an accelerator. Now we will use a quantum processor as an accelerator. They were thinking, at least on a diagrammatic level, some kind of hybrid system. But to me that raises another question, whether it's the physical architecture, especially if you're going to use a quantum processor at cryogenic temperatures, it's going to be physically separate from other processors.

A harder question, maybe quantum computers allow you to ask questions differently, right? They are not direct replacements for classical computers based on digital binary. It's liberating because you can do new things, but it's also very difficult because the whole way we think about computer science is based on numbers. Now all of a sudden you have to ask the question in a completely different way, a probabilistic way, and it seems to me that you can't just mix a classical computer and a quantum computer. In the long run, if what you're going to do is some kind of hybrid architecture, there are a lot of tough computer science problems.

HPCwire: The prevailing view in HPC seems to be that quantum computers will be a kind of accelerator for special problems. But back to the time to achieve quantum supremacy. IBM has said 2023 will be the year it delivers quantum advantage with a system of about 1,000 qubits. What do you think?

Merzbacher: I guess it depends on what the end point is. IBM has published what you call a roadmap. What will this bring? For example, I don't think anyone thought this would be a computer to crack encryption. Of course, it will have some capabilities. But whether it will prove powerful enough to be a disruptive practical use for the chemical industry or people with computational problems is unclear. I think now is a good time for IBM and others to explore these opportunities.

I've also heard that the Cloud Security Alliance (CSA) has been thinking more about security recently, releasing something saying Y2Q - the moment when we need new encryption around 2030 or 2031. So they're still almost a decade away from thinking that really powerful computers that can crack encryption. But it's faster than you might think, as it takes years to migrate to the new encryption standard.

HPCwire: Of course, NIST has its post-quantum program to help develop public-key encryption standards that can prevent quantum attacks.

Merzbacher: Right. The program has been in the works for some time and progress is being made. I think they have a goal of choosing a new standard in a year or so. That's a bit timely, considering that banks, critical infrastructure, and all the different systems take a long time to adjust. They will all have to migrate to a new encryption standard, which will take quite some time. And I think it's a long-term job because your embedded systems are not easy to upgrade, they are very important for a safe world. We are very dependent on the ability to send and receive information.
 
HPCwire: When do you think we'll see some non-security-oriented applications in real-world use? Some quantum companies say all they need is to be able to generate some "better" random numbers and selectively inject them into algorithms done on traditional systems for better results.
 
Merzbacher: I wouldn't be surprised. I wouldn't be surprised. I feel like I have to put a big disclaimer in front of everything. But as you pointed out, the intensity of the activity is very high right now. I don't know if you went to the Q2B meeting. They held their first face-to-face meeting last December. Every year in Santa Clara, it's a family gathering for quantum computing. They are very positive and optimistic, but it looks like they are more concrete and are making real progress, not just fanciful or hypothetical. It's starting to feel like in that three-year time frame, the real apps, products and features will be at hand.
 
HPCwire: You mentioned standards and benchmarking efforts for QED-C. How will the rest of the world know what you're doing?
 
Merzbacher: So, a few things. One is that we're working on some benchmarks and standards; we're not a standards-setting organization. But these discussions need to be more open and inclusive. I mean, that's why the standards development organization is very inclusive. We put our benchmarking tool on GitHub so anyone can use it. QED-C is inclusive to some extent. We welcome members from the United States, and we just recently opened our doors to members from select countries that are America's closest allies. We are now open to companies from the UK, Australia, Japan, Nordic countries, Netherlands, Canada.
 
I just came back from a trip to Europe and people are interested in QED-C for many reasons. The first is: there is definitely exciting discovery research around the world. Places like the UK are indeed hotbeds for quantum R&D. They own all these startups. These companies want to get into big markets, which, of course, means the US and elsewhere want to be part of Q-EDC so they can reach customers and take their own lifeline. So there is a lot of interest in cooperation between these like-minded countries and regions, even though they may not agree on everything.
 
HPCwire: What are you doing in the near future? What will QED-C focus on?
 
Merzbacher: One is the supply chain, which needs to be better understood. But it's not something off the shelf. We're taking some time and collaborating with others to better understand the global supply chain and the ecosystem as a whole, because if there seems to be a weakness or a gap, you can strengthen it that way.
 
We are constantly working to understand the workforce needs of this industry. There is a lack of skilled workers, especially technicians, not only with advanced degrees, but all the way down to the people you want in the lab. The COVID-19 pandemic has been devastating. I got a call today and someone said, "People want to work from home right now, sorry, but I need someone to come to the lab to do stuff." There's a shortage of that kind of staff.
 
We're trying to educate students that you don't necessarily have to get a PhD in physics; if you're a software engineer, or if you're an optics guy, there's a lot of opportunity, maybe you need to take a class or something like that, But the skills required are varied. This is an area where we continue to work on connecting people even on specific opportunities; tell me what kind of intern they need and I'll connect you with a qualified student.
 
So the supply chain, the workforce, the whole benchmarking and figuring out what's needed. The word "standard" isn't really appropriate. But we're not really at the stage where we need the interoperability standards themselves, and we need agreed-upon specifications, metrics and benchmarks. This is really the stage where we use a lot of different technologies. You mentioned single-photon sources, which is an example. We hear the same thing about low temperatures. There is no real understanding of how certain materials behave at low temperatures. You're going to put something in millikelvins and expect it to perform. Well, we need to understand how these materials behave at these temperatures. Some of these are fundamental materials problems that DOE labs or universities can really solve. We're trying to connect people who can answer these questions with people who ask them.
 
HPCwire: Thank you for your time.
 
Link:
https://www.hpcwire.com/2022/03/29/merzbacher-qa-deep-dive-into-the-quantum-economic-development-consortium/