Building the quantum information sciences (QIS) industry or more accurately, helping it build itself is Celia Merzbachers job as executive director of the Quantum Economic Development Consortium (QED-C). The QED-C was brought into being with the 2018 National Quantum Initiative Act and is broadly overseen by NIST and the National Quantum Coordination Office.
Merzbacher is a scientist (materials and nanotechnology) who brings a solid mix of research, commercial and government experience to her position. Its the perfect job for me, she says. At the Office of Science and Technology she oversaw National Nanotechnology Initiative. She also served as executive director of the Presidents Council of Advisors on Science and Technology (PCAST). At Oak Ridge National Laboratory, she was director of strategic and institutional planning. At the Semiconductor Research Corporation, she was vice president for innovative partnerships. Last year, Merzbacher was elected as a AAAS Fellow.
Fostering development of quantum computing is only part of QED-Cs mission. Understanding the technology and commercial needs of the entire quantum landscape and accelerating its progress is the broad goal. Recently, Merzbacher talked with HPCwire about the challenges and opportunities. In the course of our conversation, she noted: [You] asked at the beginning a question around what does this whole landscape look like? You would think if anybody had that map, it would be QED-C. But its not something that is readily available.
Presented here are portions of that interview, which covered technology bottlenecks, workforce issues, governance and coordination challenges, and QED-Cs near-term priorities.
HPCwire: Lets start with some background on QED-C.
Merzbacher: First, we arent just about computing. Were really about all applications of Quantum 2.0, as its sometimes called. The organization was actually called for in the 2018 National Quantum Initiative, which said NIST was to establish a consortium of stakeholders. So thats QED-C. We got what I call startup funding from NIST, but were really industry-driven. We have today about 200 non-government members, predominantly corporate, so about 150-plus companies of all sizes from all parts of the quantum ecosystem or, if you want to call it, the supply chain. That includes tier-two and tier-three suppliers of measurement capabilities or equipment or lasers or electronics and other non-quantum but enabling technologies, all the way up to system developers and software and application developers and even end users who I would say are smartly looking at how quantum technology is advancing and how it might intersect their own future in a sort of disruptive way.
You can see all of our current members on our website. Its predominantly in terms of numbers of companies, small- and medium- sized companies, although there are all the big players you would think of: IBM, Google, Microsoft, Amazon, etc. Of course, those companies arent pure-play quantum companies; they sort of have a little quantum startup inside of them. So all of our members behave a little bit like a small company in some sense. We also have companies like Deloitte and Booz Allen and those kinds of consulting companies.
HPCwire: The quantum ecosystem has dramatically expanded in recent years. Whats the QED-C mission? And how do you sort of operationalize that mission?
Merzbacher: Our mission easy to say its to enable and grow the quantum industry. QED-C is a very bottom-up type organization. Its very lean and we have a number of committees where the members come together to do activities. These range from committees on use cases and understanding what the markets and what the applications are to what the size of those markets might be, what the readiness level is of the technology, and how long its going to take to get there. Collectively, we have a lot of intelligence on where things are and where theyre going, but we dont have all the answers yet. Internally, theres a lot of sort of information being shared. I hope that over time we will be able to put out more digested documentation to help get solid, credible information to the public and to the community who sometimes thinks: Is this just hype? I dont get it yet. Wheres it going?
Were in a position to help educate and explain. Over time, I hope we can do that with the sort of use cases work that were doing. We recently, for instance, had a workshop that hasnt produced a report quite yet on quantum computing for the electric grid. So thats interesting use case; lets dig in on it a little bit and bring together to that community, which maybe isnt very expert at quantum, and look at how quantum might be useful in their sector. Theres a committee focused on enabling technologies and weve done deep dives and some roadmaps on technologies in cryogenic space, in the laser space, in the electronic space, in a lot of different non-quantum technologies that have specific requirements when used in a quantum application. For example, theres not really off-the-shelf lasers or electronics that meet those specifications today. Many companies have to either test their own, modify their own, sometimes make their own. Theres a big supply chain issue, frankly, and were trying to identify those gaps so that they can be filled.
HPCwire: To your point about the supply chain point, the overlap with DOE work is interesting. I recently talked with quantum researchers at Oak Ridge National Lab and one of the things theyre doing is trying to develop more efficient single-photon sources that would be useful in quantum networks.
Merzbacher: If you go to our public website and you scroll down on the landing page, there is a button that says TACs or Technical Advisory Committees and you can see a description of all the different committees. (List of current TACs and leaders at end of article.)
HPCwire: Within the HPC community, theres a mixture of attitudes towards quantum technology. A few are excited. Most recognize the long-term potential. But some have grown a little tone-deaf to the stream of promises. Whats your sense of the bottlenecks? What do you see as the rate-limiters to progress?
Merzbacher:As you pointed out in your comments about quantums various stakeholders, the bottlenecks are kind of up and down the whole stack. So if what youre talking about is sensing capabilities for navigation systems, thats got a different set of technology bottlenecks and requirements than IBMs superconducting quantum computer. And IBMs superconducting quantum computer has different bottlenecks from Quantinuums trapped-ion system. Its really kind of a wide-open field.
The other thing that I would say is that at the same time you have all of this activity going on in the private sector, youve also got all the investment going on in the public space, the government programs and that money thats going to Oak Ridge, for instance, or NIST. So those investments need to be aware of whats going on in the private sector and what industry needs. Thats where QED-C is playing a rather important role by doing these gap analyses at our workshops where we do a deep dive. We did one on single-photon sources and detectors. Not so much on what the technology requirements are, its almost one step before that; it turns out there isnt even agreement on how to measure and characterize those sources. And Oak Ridge is not going to ever be a manufacturer of sources and detectors at any scale; theyre a research institution, theyre trying to push the boundaries of capabilities. You know, I have this bumper sticker I want to make, leave no photon behind, because you have a single photon youre tracking [and] if you have scattering or loss of one photon, you know, its a disaster.
So its a very different set of requirements from your standard telecom system. You need to really have unique ways to characterize and measure and specify the technology, and thats not even agreed upon. So a lot of what we talk about is the need for standards, but not the ultimate standards, just standard language to describe a single photon source or detector, for example. Thats a commercial problem. Thats not really a basic research issue. QED-C is really trying to connect the world of the basic researchers, and what theyre doing, which is pushing the boundaries of science, and the people who are trying to figure out how to make that into a business.
HPCwire: How does that coordination happen? As I understand it, there is a formal council for the DOEs National QIS Centers with a chair drawn from the center directors [that] rotates on an annual basis. Its role is to set priorities for the centers. But we dont have an overall Quantum Czar.
Merzbacher: Well there is Charlie Tahan in the White House. Right? Hes sitting there at a coordination office (National Quantum Coordination Office)thats interagency. It doesnt have the purse strings, he doesnt have the checkbook, but he has quite a lot of ability to make sure that that whats happening across all these different departments and agencies is being coordinated. Thats his job,
HPCwire: Good point. How does how does the QED-C govern itself?
Merzbacher: We have a steering committee made up of members who are elected by the membership. Remember, were not government. We get a modest amount of funding from the government, but we are not a government agency, we are not controlled by the government, were controlled by our members. I have 200 bosses. The steering committee has four representatives from small companies and three representatives from large companies. Its intentional that more seats are held by small companies. And there are two seats for government agencies. Today, those [latter two seats] happen to be from NIST and the Department of Energy. Thats my board, you could call it. Ill point out that QEDC is actually not even a legal thing. Its actually administered by SRI, so I work for SRI International.
HPCwire: I didnt realize SRI was involved.
Celia Merzbacher: Yes, its running the consortium on behalf of the membership and the government.
HPCwire: Relatively speaking, QED-C is still young. How do you gauge your progress? And what are your milestones for this year and next? How do you measure your success?
Merzbacher: Its challenging. In a sense, we do a lot of what trade associations do and I dont know how they measure their progress. Were not a lobbying organization thats a bright line because I work for an organization, SRI, which does a lot of government contract work. So we dont lobby but we can educate. We certainly are about educating government policymakers about whats happening on the industry side so that they can make smart decisions. That means we go and educate, you know, examiners at the Patent Office, or we meet with people who have a role in responsibility for developing export control regulations, or meet with program managers at DOD or DOE, or NSF and say, Hey, theres a lack of fundamental understanding in this area. This is the basic research that the industry would love for government to cover, in addition to all the other stuff you do.
We also do a lot to help our small member companies, just helping them to understand how, as a business, they need to be aware of things. Some of it isnt specifically quantum. [For example, if] they need to get a handle on what kind of compliance is important when youre doing government contracts, or pointing them to funding opportunities and helping to helping them find interns or summer jobs because we have universities as members too. We connect the students at the universities to the companies where there are jobs. We do a lot of different things to try to get the bottlenecks unstuck. Were also working in this area of benchmarking and standards. There are long-used benchmarks for high-performance computing. How do you measure progress in quantum computing today? We have people who are starting to do some work in that kind of space as well.
HPCwire: What reports does QED-C issue? Are you required to submit (at least to NIST) an annual report or a quarterly update? Are those public documents?
Merzbacher: We meet with NIST constantly. We have a contract between SRI and NIST, so theres reporting requirements there. The documents we deliver to NIST are not public. We do put out some materials publicly, but this is sort of a classic member-based organization. Members are paying to be members and get some benefits as a result. Some of our reports are shared only among the members and some are made public like one we issued recently, which actually might be of interest. It was on the requirements for the intermediate representation, or the abstraction layer, between hardware and software and quantum computer. We did a workshop, a deep dive on that. The report was one we decided to make public, [because] we want the whole world to be thinking about, you know, how to run lots of different software on lots of different hardware.
Thats an example of a report that we put out. We also put one out last fall called A Guide to a Quantum-Safe Organization. It was aimed more at what I call the long-suffering CIO, who has responsibility for the security of IT systems at the company; theyve heard of quantum computing, theyre not sure when its coming and what they need to be worried about. We put out a report to try to educate people in those kinds of roles about what the threats are from quantum computing and what they should be thinking about today to prepare. Thats a specific aspect of quantum computing that it has the potential to break encryption.
HPCwire: We havent talked much about when quantum information sciences will start to deliver concrete benefits. In quantum computing, the race is to achieve quantum advantage on NISQ (noisy intermediate-scale quantum) computers. Whats your take?
Merzebacher: I think quantum advantage is what everybody is excited about and eager to see and its challenging, for a number of reasons. One is certainly there are really big problems to be overcome to build a quantum computer that is powerful enough. Youve got problems in everything [like] the quality of the qubits and their fidelity and their connectedness. Then youve got problems with error correction, and environmental control and scaling. Scaling, if I had to say, is the biggest issue in the next year or two. Even if you can demonstrate something in the lab, and youve got this hero sample that is great, how do you put that into manufacturing and scale it up?
The other problem is youve got moving goalposts because existing high-performance computing keeps getting better. Its hard to project into the future when youre going to have this crossover. People at the labs, at places like Oak Ridge, who are much more expert than I, say quantum computing is going to be like an accelerator, just like graphical processors were an accelerator. Now were going to have quantum processors as accelerators. Theyre thinking, at least on the sort of schematic level, of some kind of hybrid system. But to me that [poses] another world of questions, whether its the physical architecture, especially if youre going to have the quantum processor at cryogenic temperatures, its going to be separate physically from other processors.
An even harder problem, maybe, is that a quantum computer allows you to ask questions in a different way, right? Theyre not a drop-in replacement for a digital binary-based, classical computer. Thats liberating because you can do new things but its also hugely difficult because our whole way of thinking about computer science is based on digital. Now, all of a sudden, you have to ask questions in a fundamentally different way, a probabilistic way such that you cant just hybridize a classical and quantum computer it seems to me. There are a lot of tough computer science problems if what youre planning to do is have some kind of a hybrid architecture in the long run.
HPCwire: It does seem the prevailing view in HPC is that quantum computers will become a kind of accelerator for special kinds of problems. But back to the timing of achieving quantum advantage. IBM has said 2023 will be the year it delivers quantum advantage with a ~1000-qubit system. Whats your sense?
Merzbacher: I suppose it depends on what the endpoint is. IBM has published a roadmap that youre quoting. What will that enable? I dont think that anybody thinks thats going to be the computer that will break encryption, for instance. Sure, itll have some capability. But whether it will prove sufficiently powerful to be a sort of disruptive, practical use for chemical industries, or folks who are have computationally hard problems is unclear. I think now is a great time for IBM and others to be exploring those opportunities.
I have also heard that the Cloud Security Alliance, thinking more about the security side, put something out recently that said the sort of Y2K, moment when we will need to have new encryption in place is around 2030 or 2031. So theyre thinking a really powerful computer thats able to break encryption is still be almost a decade out. But that is sooner than you think in the sense that it takes years to migrate to a new encryption standard and new cryptographic standards.
HPCwire: Of course, NIST has its Post-Quantum program to help develop public-key cryptography standardsable to prevent quantum attacks.
Merzbacher: Right. That program has been going for some time, and its making progress. I think they have a target within a year or so to select a new standard. That is sort of just in time, given how long it takes for banks and critical infrastructure and all of the different systems that are out there to adjust. They are all going to have to migrate to a new encryption standard thats going to take quite a long time. And I think of it as having a sort of a long tail, because you have embedded systems that are not readily upgraded and how important they are to a secure world. We rely so much on the ability to send and receive information.
HPCwire: Whats your sense of when we will see some applications, not security-oriented, in practical use? A few 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 to get better results. (See the HPCwire article: Zapata Computings Formula for Achieving Quantum Advantage Today)
Merzbacher: I wouldnt be surprised. I feel like I have to put one of those big disclaimer statements in front of everything. But as you pointed out, theres such an intensity of activity going on right now. I dont know if you go to the Q2B meetings. They had an in-person meeting for the first time last December. Its every year in Santa Clara and is a sort of gathering of the quantum computing clan. I felt this past year that there was a sort of solid quality to the presentations. Theyre very promotional and optimistic, of course, but it seemed like they were more concrete and real progress was being made and it wasnt just vaporware or hypotheticals. It was starting to feel like it was in that sort of three-year time horizon when real applications and products and capabilities would be in hand.
HPCwire: Youve mentioned QED-Cs standards and benchmarking efforts. How will the rest of the world get access to the kinds of work youre doing?
Merzbacher: So, a couple of things. One is were working on benchmarks and standards with a little s; were not a standard-setting organization. But those discussions need to be more open and inclusive. I mean, thats why standards development organizations are very inclusive. We put on GitHub our benchmark tool, anybody can go and use it. Q-EDC is inclusive to a point. We welcome members who are U.S.-based and we just recently opened up to members from select countries from the closest allies to the U.S. Were open now to companies from the UK, Australia, Japan, Nordic countries, the Netherlands, Canada.
I just came back from a trip to Europe and theres a lot of interest in QED-C, and its for a number of reasons. Number one is: certainly theres exciting discovery research going on worldwide. Places like the Netherlands are really hotbeds of quantum R&D as is the UK. They have all these startups happening. Those companies want to tap into the big markets and, of course, that would mean the U.S. along with other places. They all want to be part of Q-EDC so they can kind of have access to customers and have their fingers on the pulse. Then there are suppliers to my members on this side of the ocean. So theres a lot of interest in collaboration among these sort of like-minded countries and regions, even though they may not agree about everything.
HPCwire: What on your near-term to-do list going forward? What will QED-C focus on?
Merzbacher: One is the supply chain and understanding it better. You even asked at the beginning a question around what does this whole landscape look like? You would think if anybody had that map, it would be QED-C. But its not something that is readily available. We are spending some time and working with others to develop a better picture of the supply chain and the whole ecosystem globally, because that way you can strengthen it if it looks like theres weaknesses or gaps.
Weve have an ongoing effort to understand the workforce needs of this industry. Theres a lack of skilled workers, and its drifting more and more towards people who are at the technician level, not just the advanced degrees but all the way down to the folks you want in the lab. COVID has been disruptive. I was on a call today with somebody who said, people all want to work from home now, and sorry, but I need people to come in and build stuff in the lab. Theres a shortage of that kind of worker.
Were trying to get the word out to the students that you dont have to get a Ph.D. in physics; if youre a software engineer or if youre an optics person, theres lots of opportunities, maybe you need to take one class or something like that, but theres a real diversity of skills that are needed. This is an area where we continue to work and try to connect people even at the specific opportunity [level]; tell me who has an intern job theyre trying to fill in and Ill try to connect you with a student whos qualified.
So supply chain, workforce, this whole benchmarking and figuring out whats needed. Standards is not really the right word. Its sort of used a lot. But were not really at the stage where we need interoperability standards per se, and [where] we need agreed-upon specifications and metrics and benchmarks. Thats really the stage were at with a lot of different technologies. You mentioned single-photon sources. Thats an example of one. We hear the same thing about cryogenic issues. Its not really understood what the properties of certain materials are at cryogenic temperatures. Youre going to put something down at millikelvin [temperatures] and expect it to perform. Well, we need to understand how do these materials behave at those temperatures, [such as] tables and all kinds of things that are going to be expected to held at low temperatures and perform for periods of time. Some of those are the kind of fundamental materials problems that really a DOE lab or university, even, would be certainly capable of addressing. Were trying to connect the people with the ability to answer those questions with the folks who are asking them.
HPCwire: Thank you for your time.
About Celia MerzbacherDr. Celia Merzbacher is the QED-C Executive Director responsible for continuing to build the consortium and managing operational aspects. Previously, Dr. Merzbacher was Vice President for Innovative Partnerships at the Semiconductor Research Corporation, a consortium of the semiconductor industry. In 2003-2008, she was Assistant Director for Technology R&D in the White House Office of Science and Technology Policy, where she oversaw the establishment and coordination of the National Nanotechnology Initiative. She also served as Executive Director of the Presidents Council of Advisors on Science and Technology (PCAST).
Dr. Merzbacher began her career as a materials scientist at the U.S. Naval Research Laboratory in Washington D.C., where her research led to six patents and more than 50 technical publications. She has served as Chair of the National Materials and Manufacturing Board of the National Academies of Science, Engineering and Medicine, on the Board of Directors of ANSI, as well as on advisory boards of several university research centers.
Q-EDC TACS & Chairs (as of 12/21)
New Steering Committee members
New TAC leadership
Feature image: Honeywell/Continuums optical conditioning apparatus for use with its ion trap quantum computer.
Read the rest here:
Merzbacher Q&A: Deep Dive into the Quantum Economic Development Consortium - HPCwire