UCL Quantum Science and Technology Institute


UCLQ spin-out research could bring quantum advantage in years, not decades

17 August 2021

Phasecraft improves Hamiltonian simulation for near-term quantum computers by five orders of magnitude

MediaCentral Widget PlaceholderhJAGigCi


UCL and University of Bristol quantum software spin-out Phasecraft continues to advance the potential of quantum computing as it reveals significant software advances for near-term quantum computers. In Nature Communications, Phasecraft shares a promising new strategy for simulation of quantum systems which means that quantum computers could be used to simulate quantum materials models within 2-3 years.

Many people believe materials and chemistry simulation will be one of the first significant applications for quantum computing, allowing the simulation of complex materials beyond the capacity of even the most powerful supercomputers. Hamiltonian simulation algorithms specifically provide a way to simulate materials at the atomic level using quantum computers. This method is a stepping stone to electronic structure calculations and simulations of battery cathode materials, photovoltaic solar cells, and other high-tech materials and chemistry applications.

Quantum materials simulation could speed up discovery, saving years and billions in costs. Real-world impact by speeding up materials R&D Statista reports that companies spent $41 billion in 2019 for R&D in the advanced materials and chemicals sector to develop and test new materials for potential industrial use.

Professor Toby Cubitt (UCL Department of Computer Science and UCLQ), Phasecraft co-founder and co-lead of the study said: “Compared to the best previous research, our work shows an improvement of 5 orders of magnitude. Our new algorithm needs only 259 logical time steps, even accounting for noise and errors, compared to the previously recorded best of 1,243,586. This takes Hamiltonian simulation, previously believed to require fully-scalable, fault-tolerant quantum computers that is still decades off, within reach of much nearer-term hardware. While our results do not yet fit the limited capacity of current quantum hardware, which are still an order of magnitude away from being able to run algorithms requiring this run-time, it brings us significantly closer. Algorithmic improvements of this magnitude reduce the timescale for applications of quantum computers by many years. Based on quantum hardware roadmaps from leaders such as IBM and Google, we may be able to see quantum advantage with useful scientific applications on quantum hardware within 2-3 years.”

Useful applications for noisy, intermediate-scale quantum (NISQ) hardware Quantum hardware’s current capacity is far too limited to run instances of even the simplest textbook quantum algorithms -- and that’s without adding the error correction and fault tolerance required for large-scale quantum computations. Top-end quantum hardware currently allows ≈ 50 qubits capable of implementing quantum circuits up to a gate depth of ≈ 20. Real-world applications require techniques that go beyond the state-of-the-art.

Phasecraft creates real-world quantum applications, accounting for the capacity of NISQ devices. Its research includes benchmarking error mitigation and error correction techniques implemented within software algorithms to account for how quantum hardware functions in the real world.

Research co-lead Laura Clinton of Phasecraft and PhD student at UCL’s EPSRC Doctoral Training Centre in Delivering Quantum Technologies said: “The idea for the research was to figure out how we can squeeze all the usefulness out of near-term quantum hardware. Our research shows new techniques that can be used across near-term noisy, intermediate quantum (NISQ) hardware to implement algorithms more efficiently as well as reduce the propagation of errors, by exploiting ‘bare metal’ control of qubit-level interactions, below the circuit and gate level.

Research team member Johannes Bausch of Phasecraft and University of Cambridge added, “In classical computation, it is common to perform low-level optimizations for time-critical subroutines. On quantum hardware, this line of thinking is far more challenging. We had to first identify what ‘shortcuts’ quantum hardware affords, and how much these allow us to push the limits of what is possible. This involved both rigorous analytical calculations, as well as extensive numerical simulations to pinpoint exactly what would be necessary on the hardware side to run the Hamiltonian simulation algorithms we propose.”

Phasecraft co-founder Professor Ashley Montanaro (University of Bristol) added, “Phasecraft exists to demonstrate useful, real-world applications on the error-prone, noisy quantum hardware that exists now for the near term. While companies like Google (a Phasecraft partner) have recently declared that a useful quantum computer will be publicly available by 2029, Phasecraft’s work shows there’s potential to use quantum computing for useful applications beyond lab tests and demonstrations long before that deadline.”

As a next step, the Phasecraft team is working on small-scale implementations and demonstrations on industry hardware, with quantum hardware partners including Rigetti and IBM. Phasecraft is also working with industry partners, including Johnson Matthey, in a UKRI-facilitated project to look at efficient ways to use quantum computing to simulate battery materials.

Yuan Su, Research Scientist at Google Quantum AI said: “The result obtained by Clinton, Bausch, and Cubitt constitutes a paradigm shift in developing quantum algorithms for near-term quantum devices. By considering a more realistic computation model and employing a stack of techniques for error analysis and circuit implementation, they achieved impressive improvements over previous work on the quantum resource estimate for Hamiltonian simulation, bringing practical quantum computation closer to reality."

Marco Paini, Technology Partnerships Director, Europe, at Rigetti said: “We have worked with the Phasecraft team since September 2020 to build the UK’s first large-scale commercial quantum computer. We’ve been impressed with the team’s technical capability and creativity in pushing the limits of what quantum computing can do in the near term. We look forward to our continued collaboration with Phasecraft to test this new model on Rigetti hardware to accelerate quantum advantage and the pursuit of practical applications.”



Quantum Sub-circuit Speedup Animation Credit: Phasecraft