Adaptive Eigensolver Optimisation Reduces Quantum Circuit Measurement Overhead Significantly

Quantum computing promises revolutionary advances, but current ‘noisy intermediate-scale’ devices require ingenious methods to overcome limitations in processing power, and a team led by Azhar Ikhtiarudin and Gagus Ketut Sunnardianto, from the Bandung Institute of Technology and the National Research and Innovation Agency (BRIN) respectively, now presents a significant step forward. Their research addresses a key challenge in a promising technique called Adaptive Eigensolver-Variational Quantum Eigensolver (ADAPT-VQE), which typically demands a large number of measurements, or ‘shots’, to function effectively. The team, including Fadjar Fathurrahman, Mohammad Kemal Agusta, and Hermawan Kresno Dipojono, demonstrates that by cleverly reusing data from initial calculations and allocating measurement resources based on the uncertainty of the results, they can dramatically reduce the number of shots needed to achieve accurate chemical simulations, bringing practical quantum computation closer to reality. This innovative approach maintains the reliability of the calculations while significantly improving efficiency, paving the way for more complex molecular modelling on near-term quantum hardware.

ADAPT-VQE Faces Quantum Measurement Scaling Limits

The Center for Quantum Physics in Indonesia is advancing the Adaptive Variational Quantum Eigensolver (ADAPT-VQE), a promising quantum algorithm for current generation quantum computers. ADAPT-VQE simplifies circuits and eases classical optimization challenges, but requires a large number of quantum measurements, known as “shots,” limiting its scalability on existing hardware. Improving the efficiency of ADAPT-VQE by reducing measurements without compromising accuracy is crucial for practical implementation and broader use in quantum computation.

Variational Quantum Eigensolver and Hybrid Algorithms

Extensive research focuses on the Variational Quantum Eigensolver (VQE) and related hybrid quantum-classical algorithms for complex computational problems. Foundational work by McClean and colleagues in 2016 established the principles of combining quantum computation for expectation value estimation with classical optimization. Early studies, such as that by Aspuru-Guzik and colleagues in 2005, demonstrated the potential of quantum computation for simulating molecular energies. A significant area of investigation centers on minimizing the number of measurements required for VQE, a major bottleneck for near-term implementation.

Researchers have highlighted the issue of redundant measurements and proposed optimization strategies, including measuring compatible operators in a single series and grouping measurements into commuting families. Further advancements include overlapped grouping measurements and efficient measurement of Pauli operators accounting for errors. Another key focus is the development of suitable quantum circuits, known as ansatze. Hardware-efficient ansatze are designed for easy implementation on current hardware, while chemically motivated ansatze aim to better capture the essential physics of molecular systems.

Adaptive ansatze dynamically construct the circuit during optimization, potentially leading to more efficient computations. Addressing errors and noise is also crucial, with techniques like reference-state error mitigation offering ways to improve accuracy. Research also considers specific quantum hardware platforms, such as trapped ions and superconducting qubits. Integrating classical and quantum methods is a growing area of interest. This vibrant research community is actively addressing the challenges and opportunities in quantum chemistry and quantum computing, paving the way for more powerful and efficient simulations of complex systems.

Data Reuse Reduces Quantum Measurement Overhead

Researchers have improved the Adaptive Variational Quantum Eigensolver (ADAPT-VQE), a quantum algorithm designed for problems beyond the reach of conventional computers. ADAPT-VQE builds quantum circuits step-by-step, adapting to the problem at hand to minimize resources, but traditionally requires a large number of measurements, known as “shots. ” This work addresses that limitation with two key strategies, significantly reducing measurement overhead without sacrificing accuracy. The first improvement intelligently reuses data from earlier calculations. During circuit refinement, the algorithm generates measurement results; the team discovered a way to repurpose these existing results when selecting components to add to the circuit in subsequent steps, avoiding redundant calculations and streamlining the process.

This differs from previous approaches by focusing on reusing data directly from the computational basis, minimizing both quantum and classical computational costs. The second strategy focuses on optimizing how measurements are allocated. By grouping parts of the calculation that commute, the researchers can further reduce redundancy. They then apply a technique called variance-based shot allocation, which prioritizes measurements that provide the most information, ensuring each shot contributes meaningfully to the final result, extending existing allocation methods to include gradient measurements specifically for ADAPT-VQE. Combined, these methods demonstrate a substantial reduction in the number of shots required to achieve “chemical accuracy,” a level of precision crucial for simulating molecular systems. This improvement brings ADAPT-VQE closer to practical application on current, limited-capacity quantum computers, paving the way for more efficient and accurate quantum simulations of complex chemical and materials problems.

Reducing Measurement Costs in Variational Algorithms

This research presents strategies to reduce the substantial measurement costs associated with the Adaptive Eigensolver (ADAPT-VQE) algorithm, a promising approach for quantum computation in the current era of limited quantum hardware. The team successfully demonstrated that reusing data from initial quantum measurements, and intelligently allocating measurement resources based on variance, both significantly reduce the number of shots required to achieve accurate results for molecular simulations. Specifically, the reuse of Pauli measurements reduced average shot usage by over 30%, while variance-based allocation achieved reductions of up to 43% for certain methods and molecules, maintaining the fidelity of the calculations across the studied molecular systems. The authors acknowledge that implementing these methods introduces some classical computational overhead, but argue this is minimal, particularly for the Pauli measurement reuse protocol, and is a reasonable trade-off given the limitations of current quantum hardware. Future work should focus on evaluating the effectiveness of these strategies on real quantum devices with realistic noise profiles, and exploring their performance with different operator pools and larger molecular systems.