24 December 2022

DARPA’s explorations in quantum computing search for the art of the possible in the realm of the improbable

BARRY ROSENBERG

To discuss the state of quantum computing and its military applications, we talked with Joe Altepeter, a program manager in DARPA’s Defense Sciences Office (DSO). Altepeter manages two of DARPA’s three main quantum programs — including the (US2QC) program, which is about uncovering new, novel, and overlooked avenues in quantum exploration.

Breaking Defense: Quantum computing is talked about as something that can be both offensive in the sense that it has the capability to break all known encryption, and defensive to prevent adversaries from breaking US encryption. Which is the priority for the US government? Or is it both?

Joe Altepeter is program manager of DARPA’s quantum program called Underexplored Systems for Utility-Scale Quantum Computing.

Altepeter: I’m going to choose secret option number three. The interest in quantum computers took off in 1995 when Peter Shor discovered an algorithm for efficiently factoring large numbers. I’m not an encryption expert, but I don’t think that breaks all kinds of encryption, though it certainly breaks some like RSA. That’s why NIST and agencies like that are developing alternative means of encryption that are resistant to the kinds of quantum attacks that you’re talking about.

At DARPA, our mandate is to eliminate strategic surprise. While people have been thinking about quantum computers and factoring for decades now, we’re interested in the next application which might take us all by surprise and lead to a computing revolution. We’re interested in [knowing if there] are there other surprising uses of quantum computers.

If we can build a quantum computer, will it really change how we think about computing and revolutionize computing disciplines? Or will it not really do anything that a classic supercomputer couldn’t do? The corollary to that is, let’s assume it is going to be revolutionary. These computers are really hard to build. Is there a surprising path to build one that the conventional quantum-computing community might have overlooked that DARPA needs to find out [about] this path is going to work?

Of the 10 smartest physicists I know, about half of them are convinced that quantum computers are going to totally revolutionize computing in the 21st century and be a revolutionary way to solve problems from material science, to chemistry, to mathematics, to optimization. The other half are convinced that it will never do anything that a regular or classical computer won’t be able to do.

When I think about strategic surprise, it’s hard for me to think of a discipline that has more potential for surprise than one where we think that it’s somewhere between totally revolutionary and totally useless. It’s somewhere in that zone. DARPA wants to try to bring some clarity to that question.

DARPA says many physicists think quantum computing is revolutionary. Others think it won’t be much better than today’s supercomputers. Shown is a Cray supercomputer at NASA’s Lewis Research Center in 2009. (NASA.)

Breaking Defense: What is the status of quantum in the DoD now? Is it still only in the realm of DARPA and the other research agencies?

Altepeter: As far as I know, the DoD doesn’t use quantum computers for any real problems right now. However, quantum computers, and particularly the ones that are being developed in the commercial industry, have done some amazing, near miraculous stuff. It’s been shown that quantum computers can now do calculations that are totally impossible for any classical computer anywhere on earth to do.

You and your readers might be thinking, that certainly sounds like it’s useful that they can do something that’s totally impossible for anything else to do. At this point, quantum computers are being used for proof-of-principle problems that show we can do something that nothing else can do, but we haven’t taken that next step of bending that computing power to a useful problem that we really need an answer to.

For the DoD, that means having an effect in the field, doing something we care about. There’s a lot of tantalizing pathways. If we keep getting better quantum computers and we better understand how to bend them against the problems we care about, maybe we can solve corrosion resistance and save the Navy billions of dollars maintaining their ships. Maybe we can come up with better pharmaceuticals and significantly reduce the cost of the drugs we need to care for our people. Maybe we can solve optimization problems and significantly increase the efficiency of everything we’re doing.

That said, from my perspective, we don’t have a clear path of how to solve those yet. The reason that DARPA is stepping in is to try to help get clarity on what is the link and how hard is it going to be, if we can at all, connect the really miraculous machines that we have and we’ll have in the next few years with the problems that the DoD really cares about.

Breaking Defense: When I read the background on DARPA’s three quantum computing programs, I was particularly interested in US2QC because of its mission to speed up quantum development through unexplored avenues. Describe the program.

Altepeter: There’s a lot of hype in this space, which is understandable because people are excited about the potential for this technology. But it makes it hard to tell what’s for real and what isn’t. Particularly when there’s a lot of commercial companies pursuing this technology, dozens of them.

Understandably, a lot of the secret sauce, the things that make their approaches work, are kept as trade secrets. But DARPA is interested in understanding approaches which are different from things that we’ve pursued in the past.

We put out a call that said: if you’re a company, university, or organization and think you have found the solution to build a big, powerful quantum computer and you think you’re on the path to do so, we would like to give you an opportunity to prove it. We’ll put together a validation team of some of the best experts in and around government and work with you to give you enough funding so that you won’t be slowed down.

We can provide a lot of value to these organizations by being a neutral third party to ask hard questions. We think that they can provide value to DARPA in its primary mission, which is to avoid being surprised if there really is a fantastic route out there that doesn’t look like what we’ve tried before to get to a working quantum computer.

Breaking Defense: Through what means are you trying to speed up quantum development — software, chips, AI?

Altepeter: Up until now, for the past 20 years or so, many organizations [though] not all of them, have been focused on trying to prove that this isn’t all just the realm of imagination, that it really is possible to build quantum computers that can do things that are impossible for any other computer to do. We’ve met that milestone as a society.

The biggest next step is to start focusing on the end goal. Instead of focusing on whether it is possible to make next year’s computers twice as quantum as this year’s computers, we need to look ahead. Maybe it’s decades away, but what computational capability would be a game changer for the DoD, for the commercial space? What would help us fight climate change? What would make the world a better place?

Focusing on that goal and not just saying that it would really be great if we had better batteries or didn’t have to worry about corrosion on Navy ships. But instead saying, “I want this specific computational capability to calculate the structure of this molecule or this material at this scale with these parameters.” If I could do that, I’d have a much better chance of inventing something that’s going to help the DoD in the field, in real life.

If that’s our goal, if we’re trying to get to the moon, we have to stop measuring progress by how high the planes are flying. We’ve got to figure out what’s the R&D path, what are the metrics, what’s the plan that’s really going to get us there?

From my perspective, it’s not a particular chip, it’s not a particular type of interconnect, it’s not getting a CPU speed up to a certain amount. It’s focusing on our shared goal and then deriving what are the most important pieces that we want to make a DARPA-style push to improve.

Breaking Defense: Final thoughts?

Altepeter: DARPA is not convinced that quantum is going to be a revolutionary capability. We’re also not a skeptic who thinks that quantum’s definitely not going to work. We want to go in with a clear-eyed view and do a rigorous evaluation to see where and how and on what path quantum can make the DoD’s capabilities better, and can really make the world a better place. That’s what we’re trying to do, not prejudge the outcome, but take a hard look and see what we learn and reduce strategic surprise in this space.

No comments: