Researchers Bound Quantum Eigensolver Shots to Linear Scaling

Scientists are developing quantum subspace diagonalization methods as promising algorithms for quantum chemistry on near-term quantum computers. These methods can estimate low-lying energies of molecular systems using shallow quantum circuits. This estimation requires many circuit repetitions to determine the projection. Scalable thresholding reduces quantum computational cost for molecular energy simulations A significant reduction in the per-matrix-element shot count needed for accurate quantum calculations has been achieved, improving the scaling from O(M³) to O(M), where M represents the number of reference states. This represents an advancement in the field of quantum computational chemistry, as the computational cost associated with determining molecular energies has historically been a major impediment to simulating larger systems. The improvement hinges on a scalable thresholding scheme that selectively removes poorly overlapping reference states from the calculation. Previously, identifying and discarding these states was computationally expensive, often negating any potential benefits. The core principle behind this technique lies in recognising that reference states exhibiting minimal overlap contribute disproportionately to the noise and instability of the calculation, without significantly impacting the accuracy of the final energy estimate. By judiciously eliminating these states, the computational effort can be dramatically reduced. Establishing that the sensitivity of eigenvalues is now governed by the condition number of the remaining overlap matrix, rather than the initial, larger dimension, represents a step forward in controlling the error propagation within the algorithm. The condition number of a matrix is a measure of its sensitivity to perturbations; a high condition number indicates that even small errors in the input data can lead to large errors in the output. By ensuring that the condition number remains manageable, researchers can effectively limit the impact of noise and imperfections inherent in near-term quantum hardware. This is particularly crucial for algorithms like the nonorthogonal quantum eigensolver (NOQE), which are susceptible to errors in the estimation of matrix elements. The thresholding process effectively ‘cleans’ the subspace, leading to a more stable and accurate eigenvalue solution. Hydrogen chains and rings were used as model systems in simulations, employing the PySCF and OpenFermion software packages alongside AI coding assistants to rigorously verify the improved scaling behaviour. These software packages provide robust tools for performing quantum chemical calculations and for translating these calculations into quantum circuits suitable for execution on quantum computers. The simulations were performed with the STO-3G basis set, a minimal basis set commonly used for initial calculations, and interatomic distances were fixed at 1.5 Angstroms to simplify the analysis. The choice of STO-3G allows for faster computation, enabling the exploration of larger systems, although more accurate results would be obtained with larger basis sets. The results show a direct correlation between the number of reference states, denoted as M, and the number of hydrogen atoms in the system, confirming the linear scaling of the computational cost with system size. Analysis of the condition number revealed successful control of this value, preventing its rapid escalation with system size, and thus maintaining the stability of the calculation. Furthermore, spin contamination in the NOQE method was sharply reduced compared to non-orthogonal configuration interaction (NOCI), indicating improved accuracy and highlighting the potential for more reliable simulations. Spin contamination arises from the use of non-orthogonal reference states and can lead to inaccurate energy estimates; the reduction in spin contamination observed in the NOQE method suggests that the thresholding scheme is effectively mitigating this issue. The optimal threshold is not universal and could inadvertently remove important information, a point requiring future investigation, as the ideal balance between computational cost and accuracy needs to be carefully determined for different molecular systems and basis sets. Linear scaling of measurements improves molecular simulation feasibility Minimising the resources needed to achieve accurate results is vital when simulating molecular properties using quantum computers, particularly given the limitations of current hardware. The nonorthogonal quantum eigensolver offers a promising path towards achieving this goal, but its effectiveness relies on efficiently managing the number of measurements taken from the quantum processor. Each measurement introduces noise and uncertainty, and the total number of measurements required to obtain a reliable result can quickly become prohibitive for larger systems. A major advance has been demonstrated in reducing the computational burden for simulating molecules on quantum computers, despite the acknowledged limitations of the thresholding technique; the technique is not a panacea, and careful consideration must be given to the choice of threshold value. Selecting reference states carefully during calculations allows the number of measurements needed to scale linearly with the number of reference states. This is a crucial improvement, as it means that the computational cost does not increase exponentially with system size. The underlying principle is that by focusing on the most important reference states, the algorithm can achieve the same level of accuracy with significantly fewer measurements. The findings establish a clearer relationship between measurement count and the size of the quantum calculation within the nonorthogonal quantum eigensolver, providing a more predictable and manageable scaling behaviour. This predictability is essential for planning and executing quantum simulations on limited hardware resources. This improvement moves beyond simply minimising errors; it fundamentally alters how measurement demand scales with system size, offering a pathway to more efficient quantum chemistry and enabling the simulation of larger, more complex molecular systems, such as those found in materials science and drug discovery. The ability to simulate these systems with greater accuracy and efficiency could lead to breakthroughs in these fields. The linear scaling achieved represents a step towards realising the full potential of quantum computers for solving challenging problems in chemistry and materials science. The research demonstrated that, after applying a thresholding technique, the number of measurements needed to calculate molecular energies using the nonorthogonal quantum eigensolver scales linearly with the number of reference states. This represents an improvement over previous methods where measurement costs increased much more rapidly. By controlling the sensitivity of eigenvalues through the condition of the overlap matrix, researchers achieved a more predictable scaling behaviour for quantum calculations. The authors suggest this approach could enable simulations of larger molecular systems, although careful selection of the threshold value remains important. 👉 More information🗞 Improved Measurement Cost Scaling in the Nonorthogonal Quantum Eigensolver✍️ Mingyu Kang and K. Birgitta Whaley🧠 ArXiv: https://arxiv.org/abs/2608.12830 Stay currentSee today’s quantum computing news on Quantum Zeitgeist for the latest breakthroughs in qubits, hardware, algorithms, and industry deals. Tags: