Caltech: Caltech and Amazon Partner to Create New Hub of Quantum Computing

“Most people know of the company Amazon. We can do a lot of shopping there, but it’s also a powerhouse in engineering,” says Fernando Brandão, Bren Professor of Theoretical Physics at Caltech and head of quantum algorithms at AWS. “Amazon Web Services is the biggest cloud provider on Earth today. They are thinking about how they can make computing simpler and better for people who use AWS. And they’re also thinking about the next step, and how to do computing and cloud computing in the next five or 10 years.”

The collaboration will help bridge the commercial side of quantum computing with the fundamental research taking place at Caltech, which has a long history of breakthroughs in the quantum sciences.

“AWS will benefit from the ideas percolating here on campus,” says Oskar Painter (MS ’95, PhD ’01), John G. Braun Professor of Applied Physics and Physics at Caltech and head of quantum hardware at AWS. Painter says quantum computing is still a very young technology, so it is crucial for development efforts to be directly connected to the latest research in academia.

“If we were to just take today’s ideas and go forward with them, we would create a dinosaur of a quantum computer,” Painter says. “We need to be closely connected and tied into these basic research efforts.”

This is the first corporate-partnership building on the Caltech campus, and it reflects Caltech’s interests in bringing fundamental science to the marketplace. Through scholarships, internships, and seminars, the center will also support Caltech students and early career scientists.

“Students will have the opportunity to interact with cutting-edge research by way of the center being located at Caltech. This will be pretty amazing for students,” says Painter. “And AWS can tap into that talent. Those are the future engineers and scientists who are going to build quantum computers.”

Scaling Up

One of the biggest challenges in building quantum computers is scaling them up. Because the technology behind the computers is so complex, current prototypes are still in experimental stages. For quantum computers to truly exceed what can be done with today’s classical computers—a milestone referred to as quantum advantage—they will need to be much larger.

For instance, today’s rudimentary quantum computers work on only a few dozen qubits—the quantum equivalent of bits, or the 1’s and 0’s that make up the language of classical computers. Researchers want to build quantum computers with thousands of qubits and more.

“Classical computers have billions and even trillions of bits, and that’s where we eventually would want to be with qubits,” says Brandão.

Painter says even though some media reports have suggested quantum computers are around the corner, the technology is still in its infancy. “We can do small problems now with quantum computers but we need to scale up the technology by many orders of magnitude before we can truly tackle problems with a big impact. Figuring out what problems are best solved with quantum computers is also an active area of research. There’s excitement because we are starting to be able to control large-scale quantum systems, but we don’t have all the answers yet.”

Unlike bits, qubits can exist in a quantum state known as superposition where they are both a 1 and 0 at the same time, and all the possible states in between. (In Erwin Schrödinger’s famous superposition analogy, a cat can be both dead and alive at the same time, but the cat can also be in any combination, or superposition, of those two states.)

With Power Comes Fragility

The ability of qubits to take on multiple states at once is what gives quantum computers the potential to be exponentially more powerful than classical computers for certain types of problems, including those in chemistry, finance, cryptography, and more. But that power comes with drawbacks. Qubits are very fragile; any slight perturbations, such as vibrations or heat, can nudge them out of superposition, a phenomenon known as decoherence. The key to building successful quantum computers of the future lies in controlling these errors.

“Transistors in our modern computers experience extremely low error rates at the level of one error per billion billion operations, enabling complex computations to be performed,” says Painter. “Quantum computers as of now are limited by error rates at the level of approximately one error in every thousand operations.”

A primary goal of AWS is to create a computer architecture that builds quantum error correction into the hardware. The AWS hardware relies on superconducting qubits, which operate at ultra-cold temperatures slightly above absolute zero. Quantum error correction methods use redundant sets of qubits at the physical hardware level (“physical” qubits) to form “logical” qubits, which encode the quantum information and can be used to detect and correct errors. A primary challenge in quantum error correction is the large amount of hardware associated with the number of physical qubits required per logical qubit.

“In error-corrected quantum computers, the more physical qubits that are used to make up a logical qubit, the better one can reduce the logical qubit error rate relative to the individual physical qubit error rate,” says Painter. “Going forward, we want to scale up the number of logical qubits to hundreds or thousands, while also driving the logical qubit error rate down by multiple orders of magnitude, so that we can perform quantum computations of sufficient complexity to tackle high-value problems. In order to do so, we need to further develop both the physical hardware and the logical qubit architecture.”

Quantum Roots

Caltech is well suited as a hub of quantum innovation thanks to its rich history in the field. The late Richard Feynman, a longtime Caltech physics professor, was one of the first to propose quantum computers. In a 1981 lecture, he famously explained that there are limits to simulating systems in physics with classical computers because “nature isn’t classical, dammit, and if you want to make a simulation of nature, you’d better make it quantum mechanical, and by golly, it’s a wonderful problem because it doesn’t look so easy.”

In 1994, Caltech alumnus Peter Shor (BS ’81), then at Bell Labs, developed a quantum algorithm that can factor large numbers in very short periods of time, demonstrating the enormous power of the future technology. For instance, a quantum computer would be able to factor a 2,048-digit number in eight hours, whereas this would take a classical computer about 300 trillion years. “When I heard about this, I was awestruck,” recalled Amazon scholar John Preskill, the Richard P. Feynman Professor of Theoretical Physics and the director of the Institute for Quantum Science and Technology (IQIM), in a 2013 Caltech article. Shor also helped pioneer the development of quantum error-correction codes.

Caltech’s Jeff Kimble, the William L. Valentine Professor of Physics, Emeritus, was among the first, in 1998, to achieve quantum teleportation, whereby information was sent from one light beam to another via entanglement, a process in which particles are connected without being in direct contact with each other. In 2008, he and his colleagues were also the first to store entangled quantum states in a memory device.

“I came to Caltech as a graduate student to study with people like Jeff Kimble, who was measuring small atom-photon quantum systems with exquisite sensitivity, and who was developing quantum protocols with these systems that might ultimately be used to create a form of quantum internet one day,” says Painter.

Annika Dugad, a current graduate student at Caltech, says having the AWS center on campus brings even more of a buzz to the cutting-edge quantum research happening at Caltech. Dugad is among several scholars whose graduate studies are being funded by Amazon. She is using the support to study the black-hole information paradox, a mystery first illuminated by Stephen Hawking in the early 1970s that asks what happens to information that gets trapped inside a black hole. Ultimately, Dugad says she would like to take what she has learned and apply it to more practical questions in quantum computing.

“Not many campuses have this type of quantum center,” she says. “I know there’s an outlet where I can go to collaborate on experiments. Academia is slower paced by nature, but in industry, they have deadlines and can really get things going.”

“There’s a new paradigm in computing,” says Brandão. “It’s not just about making our current computers a little bit faster or a little bit better as we have been seeing in the last 50 years at least. It’s about building a completely new kind of computer. Quantum computing is very early days, but I think it’s also very exciting.”