Researchers Cut Variational Quantum Eigensolver Energy Errors by Orders of Magnitude

Energy errors in quantum simulations have been reduced by orders of magnitude through an error-mitigation strategy called extended random phase approximation (ERPA). The Heyrovský Institute team developed ERPA to reconstruct key data, the two-particle reduced density matrix, from noisy measurements; this filters out errors improving calculations for molecular modelling. Unlike existing methods, it requires fewer measurements making it an efficient route towards reliable electronic structure calculations on near-term quantum computers. A new technique reduces errors in quantum computer simulations and improves predictions of molecular properties without requiring extensive measurements.

The method reconstructs key data, called the two-particle reduced density matrix, filtering out errors that arise during calculations of molecules using noisy quantum hardware. By combining this with classical computation, scientists are closer to reliably simulating complex chemical behaviour vital for designing materials and understanding reactions.

The Heyrovský Institute team has developed an error-reduction technique for quantum computer simulations that promises more accurate predictions of molecular behaviour without demanding excessive measurements; this builds upon existing methods like variational quantum eigensolver (VQE), where a quantum computer performs complex calculations guided by classical computation, akin to a collaborative project with defined roles.

Their approach centres around reconstructing what’s called the two-particle reduced density matrix, or 2-RDM, which provides a simplified description of how electrons interact within a molecule focusing on pairs rather than tracking each individually, much like taking notes only about group work instead of individual contributions. By filtering out errors during these key calculations using extended random phase approximation (ERPA) and combining it with established techniques such as density matrix renormalization group (DMRG), they have achieved sharp reductions in energy calculation errors; but can this method consistently deliver reliable results across increasingly complex molecular systems.

Extended Random Phase Approximation Sharply Reduces Quantum Simulation Errors

Energy errors in quantum simulations dropped by orders of magnitude using extended random phase approximation (ERPA), a novel error mitigation strategy developed collaboratively by scientists at the Academy of Sciences and Heyrovský Institute alongside Łódź University of Technology, Palacký University and Technical University; this reduction surpasses previous capabilities. ERPA reconstructs the two-particle reduced density matrix from noisy measurement results, effectively filtering out calculation inaccuracies for molecular modelling without extensive data collection, a significant advantage over methods like quantum subspace expansion which demand more measurements.

Benchmarks on water double-bond stretching revealed that these improvements reduce errors beyond levels achievable with existing techniques. Combining ERPA with multireference adiabatic connection corrections allowed recovery of ‘dynamical correlation’, yielding accurate potential energy curves and reaction barriers vital to understanding chemical processes.

Extended Random Phase Approximation Mitigates Errors in Variational Quantum Eigensolver Simulations

A new extended random phase approximation (ERPA) method intelligently filters inaccuracies arising from noisy quantum hardware measurements to reconstruct the two-particle reduced density matrix, essential for molecular modelling. Unlike quantum subspace expansion, ERPA requires only one- and two-particle RDMs for reconstruction. A hybrid scheme incorporating DMRG calculations further refined this approach; DMRG provides a noise-free approximation to the input 1-RDM used initially within their methodology.

Benchmarks focused on water double-bond stretching and cyclobutadiene automerization demonstrated substantial improvements in energy error reduction when combined with multireference adiabatic connection corrections accounting for dynamical correlation beyond the active space. While demonstrating major progress, these benchmarks are limited to just two examples, leaving open questions about its applicability to larger or more complex molecules. The use of AC0 corrections introduces additional computational cost, though whether it is necessary across all applications remains unproven by this work.

Extended Random Phase Approximation mitigates VQE error through reconstructed electron correlation functions

Extended random phase approximation (ERPA) advances error mitigation within variational quantum eigensolver calculations; it reconstructs the two-particle reduced density matrix, effectively describing how pairs of electrons behave, from noisy measurements. Benchmarks performed on quantum hardware demonstrated substantial improvements: simulating water double-bond stretching and cyclobutadiene automerization resulted in energy calculation errors being reduced by orders of magnitude.

The team also employed multireference adiabatic connection (AC0) corrections to capture dynamical correlation effects, yielding more accurate potential energy curves and reaction barriers. Further work will be needed to establish ERPA’s broad applicability with larger or chemically different systems, as scientists acknowledge its success is currently limited to these specific cases. This approach preserves important input information and allows integration with classical DMRG calculations for further refinement, creating an efficient hybrid strategy applicable on current hardware.

Extended random phase approximation (ERPA) improves error mitigation within variational quantum eigensolver calculations by reconstructing the two-particle reduced density matrix from noisy measurements. This reconstruction filters errors inherent in quantum hardware while requiring only one- and two-particle RDMs, simplifying computation compared to other methods. The authors note further research is needed to determine if this approach scales effectively to larger or more complex molecules.

More like this

Quantum AlgorithmsResearchers Calculate Disordered System Energies Using Quantum Algorithm

Quantum AlgorithmsResearchers Find Repulsive Polaron Interactions in Strong Regimes

Quantum Research NewsQuantum data re-uploading works with single-qubit measurements

Quantum PhysicsGeometric quantum state preparation solves Helmholtz equation