Christopher Monroe, University of Maryland, gave the first keynote address. He shared background information on quantum science and computing, discussed applications and potential models for quantum computing, and described the role of the National Quantum Initiative (NQI).
Quantum Science and Quantum Computing
Quantum science features radical concepts foreign to all other fields of science, such as the necessity of probabilities or that measurements can inherently affect a physical system. Quantum technology is a confluence of multiple disciplines, including physics, information theory, computer science, mathematics, chemistry, and engineering. As Richard Feynman noted in the 1950s, the laws of physics are different when working at the scale of simple systems like atoms, and this presents new opportunities, like quantum technology, but also new challenges. Today, quantum’s greatest challenge is to move from an understanding of the basic science to the creation of useful devices.
The computational power of classical computing technology is saturating Moore’s law (the exponential growth of transistor density in conventional computer processors). Quantum computing represents a major opportunity to overcome these limitations. It also presents a radical new way of thinking about information,
which raises cultural issues in addition to technological ones. However they are designed, quantum computers will be more different from today’s computers than today’s computers are from an abacus.
Quantum bits (“qubits”) are different from traditional computing bits. They can be in both the 0 and 1 state at the same time (known as a superposition) with arbitrary weightings, but only when they are perfectly isolated and not being observed or measured. When the system is observed or measured, it changes, producing a random result with probabilities given by the weightings. While scientists do not have a unified understanding of why this is and there are many interpretations, these rules have been tested to more accuracy than any other theory in nature, Monroe said.
Qubits can also be entangled with one another. Taken individually, they can be random, but when qubits are entangled and measured together, they can have definite correlations, such as always being the same state 00 or 11, but never 01 or 10. With n qubits, there are an exponential number of states possible, but measuring every qubit will result in just a single n-bit number, at random. Therefore, scientists must use “quantum gates” to collapse these probabilities to much smaller numbers of possibilities (or perhaps just a single one) and thereby produce actionable information.
Quantum Applications and Quantum Computers
The first well-known quantum computing application was the factoring of extremely large numbers, proposed by Peter Shor in the 1990s. Shor’s algorithm was immediately recognized as a threat to data encryption, prompting theoretical academic work to revise information theory to incorporate quantum superposition and entanglement.
The realm of quantum applications has expanded over the past decade. Most are related to optimization, which has applications across multiple business and academic sectors, including logistics, operations, decision making, pattern recognition, machine learning (ML), and artificial intelligence (AI). In particular, quantum computing may be well suited for solving extremely complex problems, which are inherently quantum mechanical. In materials science, minimizing complex functions, and potentially solving big data models more efficiently, without the high-performance bandwidth that classical computing requires is what is really needed to move the discovery of new materials effectively forward.
Researchers are experimenting with multiple physical platforms for quantum computing, and quantum computers made from different components may lend themselves to very different applications. Two relatively advanced approaches (see Figure 2.1) are based on fundamentally different structures: superconducting circuits and trapped atomic ions. Among leading quantum computing technologies,
these two designs have the lowest error-per-gate-per-operation rates, although they still need to improve by orders of magnitude to find practical use.
Superconducting circuits have a fast clock speed and printable circuits with the potential for very-large-scale integration (VLSI), but they have limited connectivity, are not reconfigurable, and require super-low temperatures. Trapped atomic ions can be fully connected and reconfigurable, but they have slow clock speed and require more engineering. Trapped ions can also be moved in space to multiplex qubits, and they can be manipulated with lasers, although such architectures will require continual refinement of the classical control systems (e.g., trap electrode management and optical delivery). Researchers are also considering semiconductor, neutral atom, and photonic designs.1
Monroe’s university laboratory and start-up company IonQ combine commercial components such as lasers and modulators to develop ion trapping system technology, see Figure 2.2. The group’s goal is to create a quantum computing system that can be used even by those without an understanding of how ion trapping works, similar to how anyone can use a smartphone even without knowing how it works. The IonQ team deliberately includes non-engineers and non-quantum experts; Monroe posited that this type of collaboration among many people with
1 G. Popkin, 2016, Scientists are close to building a quantum computer that can beat a conventional one, Science 354:1090.
different areas of expertise, which can be difficult to replicate in an academic setting, is required for advancing quantum technology. Workshop participant Joseph Broz, SRI International, noted that a workforce survey is under way to better understand which fields of expertise will best foster quantum computing advancements.
The National Quantum Initiative
Academic research may solve problems, but it does not necessarily build usable systems or products, while industry creates products, but does not focus on the fundamental research. NQI was created in late 2018 to bridge the gap between basic research at the university level and applied science at the industry level. Along with the U.S. Department of Defense (DoD), U.S. Department of Energy (DOE), and other agencies, NQI aims to coordinate efforts to both fund basic quantum research and push industry to translate it into technology products. In response to a question, Monroe added that national research laboratories can also play a role in advancing quantum technology, especially those sponsored by DoD or DOE. NQI offers support to quantum technology-focused research centers that combine expertise from academic, government, and industry research laboratories working on quantum technology, with the overall goal of keeping the United States at the forefront of the quantum technology field.
Jerry Chow, IBM Research, spoke of the challenges that quantum mechanics present, unresolved facets of quantum computing, and recent advancements within the broader quantum technology field—in particular, quantum communication.
Understanding Quantum Behavior
Chow began by noting Richard Feynman’s famous declaration that no one truly understands quantum mechanics—even those working in the field. While much about quantum mechanics is counter intuitive it is not unsolved, at a fundamental level, it is a theory that works just fine, and there has been progress toward its application in quantum technology.
All quantum technology relates to making use of two postulates: the uncertainty principle and entanglement. The uncertainty principle states that even a system in a perfectly defined state can behave randomly, encompassing both definiteness and randomness. Entanglement is a phenomenon whereby the behavior of two physically separated systems is both individually random yet strongly correlated with one another. That is, when two entangled elements are viewed separately,
they are completely random, but when they are viewed together, they are correlated in every measurable position.
Although defying our classical intuition, these two principles are intrinsic to quantum technology advancements, and will ultimately change the way scientists process information or develop new technology. This will reshape the entire technological landscape. For example, entanglement can dramatically improve image resolution in sensing, metrology, and other gravitational wave detection applications. Quantum technology can also be used to create truly secure quantum communication, which is increasingly important in our information-based society, although development and scaling will take many more years, in Chow’s view.
Recent Advances and Future Needs in Quantum Computing
A reimagining of the entire field of information processing may be far away, but the horizon for quantum computing is much closer. Fault-in quantum computing is probably 15 years away, according to Chow, although testing of different systems, such as adiabatic quantum computers and analog quantum simulators, is under way. Each design has its own unique challenges, timelines, and technological needs that require close examination.
The “holy grail” of quantum computing is fault-tolerant quantum computers. Fault-tolerant quantum computers can prove Shor’s algorithm, but only with qubits that do not produce noise or errors. Currently, researchers are in an era of noisy intermediate-scale quantum (NISQ) computers. These systems have errors and other aspects that are not fully understood, making it difficult to prove their characteristics and find applications. Nonetheless, experimenting with them will still reveal many unknowns about quantum mechanics and computing.
IBM is currently demonstrating quantum computers with 5-20 qubits; an eventual fault-tolerant quantum computer will likely have well over a million, Chow posited. (In assessing the success of quantum computers, raw qubit count is one metric, although the number of operations and errors, circuit width, and depth must also be considered, he noted.) The path toward fault tolerance is very challenging and includes multiple milestones, using more and more qubits to prove a true quantum advantage. Today’s small quantum processors require dedicated infrastructure to build, program, access, and use. While the hardware makeup of a fault-tolerant quantum computer is unknown, it is unlikely to involve merely scaling up current technologies.
Demonstrations using today’s quantum computers largely replicate work that can be done with classical computing, but they lay the groundwork for a future in which quantum computing unlocks vast new computational opportunities. For example, in chemistry, quantum computing was used to calculate energy curves for various compounds, although there was evidence of errors and noise, which
was somewhat mitigated in further experiments with the added benefit of extending the computational reach of the quantum computing devices.2,3 Another study demonstrated theoretical applications for a quantum kernel estimator.4 Quantum-enabled research, even on noisy hardware, is integral to the larger goal, Chow noted. Shallow circuits, hybrid algorithms, and error corrections are other important areas that need to be fully explored and experimented on to advance quantum computing, reduce the noise it generates, and make classical simulations more efficient.
An analog quantum simulator, another type of quantum computer, is a somewhat less controllable system that has been useful for studying chemistry, materials, and other areas, although there is still a lot of resulting noise. These simulators are not quite fully controllable quantum systems, but they provide a different perspective on how existing quantum systems could be used, what hardware could be developed, and what new materials and elements might be needed. Research with this technology has included studies of Rb atoms simulating magnetic phase transitions, trapped ions, and many-body dynamics,5,6 suggesting that quantum matter can be built once a better understanding of condensed matter physics and new materials emerges.
Last, novel quantum materials, with directed goals and proper parameters, will be great enablers and influencers for all quantum technologies, not just quantum computing, Chow said. Bilayer graphene and other layered materials, as well as superconductor–semiconductor hybrids such as InAs and InSb nanowires and semiconductor heterostructures supporting a two-dimensional electron gas (2DEG), have strong potential, although there are many unknowns about why or how they work and how they would impact existing quantum computing technologies. For example, scaling could be improved (without relying on brute-force approaches) if bulky isolator components could be replaced with miniaturized 2DEG versions. New materials could also improve the junction types currently
2 A. Kandala, A. Mezzacapo, K. Temme, M. Takita, M. Brink, J.M. Chow, and J.M. Gambetta, 2017, Hardware-efficient variational quantum eigensolver for small molecules and quantum magnets, Nature 549:242-246, https://doi.org/10.1038/nature23879.
3 A. Kandala, K. Temme, A.D. Córcoles, A. Mezzacapo, J.M. Chow, and J. Gambetta, 2019, Error mitigation extends the computational reach of a noisy quantum processor, Nature 567:491-495, https://doi.org/10.1038/s41586-019-1040-7.
4 V. Havlíček, A.D. Córcoles, K. Temme, A.W. Harrow, A. Kandala, J.M. Chow, and J.M. Gambetta, 2019, Supervised learning with quantum-enhanced feature spaces, Nature 567:209-212, https://doi.org/10.1038/s41586-019-0980-2.
5 J. Simon, W.S. Bakr. R. Ma, M.E. Tai, P.M. Preiss, and M. Greiner, 2011, Quantum simulation of antiferromagnetic spin chains in an optical lattice, Nature 472:307-312, https://doi.org/10.1038/nature09994.
6 J. Zhang, G. Pagano, P.W. Hess, A. Kypriandis, P. Becker, H. Kaplan, A.V. Gorshkov, Z.-X. Gong, and C. Monroe, 2017, Observation of a many-body dynamical phase transition with a 53-qubit quantum simulator, Nature 551:601-604, https://doi.org/10.1038/nature24654.
in use, making the fabrication process easier. In addition, transduction will play a large role in quantum communication, quantum computing, and discovering novel quantum materials.
Quantum communication, secure communication via quantum key distribution (QKD) links, relies on the quantum cloning theorem, which states that quantum information cannot be duplicated. Chinese researchers have already demonstrated satellite-to-ground QKD over 1,200 kilometers. However, many breakthroughs are still needed, such as quantum repeaters, transduction, and cloud-based quantum computing, before quantum communication can become widely adopted.
The potential payoffs are enormous. Quantum communication could, for example, enable cloud-based blind quantum computing, distributed quantum computing, and better quantum sensors. IBM researchers working to advance quantum transduction technologies are targeting readouts on the order of a microsecond; other research teams are exploring lambda-level systems, magnon-mediated transducers, electro-optical mechanical transductors, and direct electro-optic transductors. It is also important to monitor how progress will alter fabrication and implementation, and to be able to build systems with low-loss electro-optic materials, Chow noted.
Clearly, there remain multiple key challenges for quantum computation and communication that will require the work of the entire quantum community to overcome. Closing, Chow urged the community to continue to tackle long-term problems with an eye toward the distant horizon, to convene workshops to discuss future applications, and to create enabling technologies, not only to advance quantum research but also to maintain the position of the United States as a global technology leader.
A participant asked whether IBM researchers experimented with substituting different compounds. Chow said it did, and added that it also simulates different interactions. IBM views these chemistry problems as test cases that can help define protocols for understanding and mitigating noise or errors, rather than for pure innovation. In response to another question, Chow clarified that novel materials and crystalline junctions would help reduce variability in tunnel junctions.