A smaller statistical error bar need not mean a more accurate molecular energy. We evaluate noiselessly optimized molecular variational quantum eigensolver states on IBM Aachen and Pittsburgh, two 156-qubit Heron processors (provider-reported revision 3). Only energy evaluation runs on hardware, with parameters, layouts and compiled circuits frozen and dynamical decoupling disabled. A five-molecule pilot, three-molecule isolated cohort in three windows and known-state controls comprise 110 jobs and 28.2 billed quantum processing unit (QPU) minutes; the isolated cohort spans 30.2 h within a 31.7 h campaign. Raw 1024 O2 and BeH2 mean residuals exceed returned uncertainty by factors 12-32. Pittsburgh BeH2 cross-job standard deviation is 5.6 times mean returned uncertainty. The zero-noise extrapolation (ZNE) package reduces O2 absolute residuals by 49.7% on Aachen and 83.5% on Pittsburgh at about fourfold charge. On Pittsburgh, twirled readout error extinction (TREX) removes 80 millihartree (mHa) per Raw 1024 job charge versus 50 mHa for ZNE. Controls reveal qubit-124 preparation or readout errors 5.4-5.6 times calibration; 13 mitigated energy point estimates fall outside the mapped spectrum, with no Raw counterpart. These results support workload checks and repeated measurements before additional sampling.
AI Summary
The paper asks how to allocate QPU time among additional shots, error mitigation and independent repetitions when molecular-energy estimates are biased. It isolates energy evaluation from VQE optimization: noiselessly optimized parameters, physical layouts and compiled circuits are frozen, then evaluated on IBM Aachen and Pittsburgh, two 156-qubit Heron processors. A five-molecule pilot, a three-molecule cohort repeated across three temporal windows, and known-state controls comprise 110 jobs and 28.2 billed QPU minutes. Molecular workloads use only 1–5 qubits.
Comparisons with the ideal frozen-state energy separate execution residuals from active-space and ansatz errors. For isolated raw O$_2$ and BeH$_2$ evaluations at 1024 nominal shots, mean residuals are 12–32 times returned standard errors. Pittsburgh BeH$_2$ has cross-job SD 120.4 mHa versus mean returned uncertainty 21.6 mHa. Quadrupling shots approximately halves returned uncertainty without reliably reducing bias. The ZNE package reduces O$_2$ absolute mean residuals by 49.7% on Aachen and 83.5% on Pittsburgh at roughly fourfold charge. Pittsburgh TREX yields greater residual reduction per normalized charge than ZNE; Aachen reverses this ordering. Controls identify qubit-124 preparation/readout discrepancies 5.4–5.6 times calibration estimates. Thirteen mitigated energy estimates fall below mapped spectral minima; no raw estimate does.
The stated limitations include only two devices and three windows, fixed treatment order confounded with time, disabled dynamical decoupling, simplified chemistry models, and unavailable measurement circuits, bitstrings and learned mitigation responses. Controls do not distinguish preparation from assignment error. The shot-allocation model assumes stationary bias and square-root sampling scaling; the proposed diagnostic protocol lacks independent validation.
Little sign of AI-generated textIncremental
Written by gpt-6.1-sol from the paper's full text. Automated: it can be wrong.
Comments
38 pages, 15 figures. Includes supplementary information
A smaller statistical error bar need not mean a more accurate molecular energy. We evaluate noiselessly optimized molecular variational quantum eigensolver states on IBM Aachen and Pittsburgh, two 156-qubit Heron processors (provider-reported revision 3). Only energy evaluation runs on hardware, with parameters, layouts and compiled circuits frozen and dynamical decoupling disabled. A five-molecule pilot, three-molecule isolated cohort in three windows and known-state controls comprise 110 jobs and 28.2 billed quantum processing unit (QPU) minutes; the isolated cohort spans 30.2 h within a 31.7 h campaign. Raw 1024 O2 and BeH2 mean residuals exceed returned uncertainty by factors 12-32. Pittsburgh BeH2 cross-job standard deviation is 5.6 times mean returned uncertainty. The zero-noise extrapolation (ZNE) package reduces O2 absolute residuals by 49.7% on Aachen and 83.5% on Pittsburgh at about fourfold charge. On Pittsburgh, twirled readout error extinction (TREX) removes 80 millihartree (mHa) per Raw 1024 job charge versus 50 mHa for ZNE. Controls reveal qubit-124 preparation or readout errors 5.4-5.6 times calibration; 13 mitigated energy point estimates fall outside the mapped spectrum, with no Raw counterpart. These results support workload checks and repeated measurements before additional sampling.
The paper asks how to allocate QPU time among additional shots, error mitigation and independent repetitions when molecular-energy estimates are biased. It isolates energy evaluation from VQE optimization: noiselessly optimized parameters, physical layouts and compiled circuits are frozen, then evaluated on IBM Aachen and Pittsburgh, two 156-qubit Heron processors. A five-molecule pilot, a three-molecule cohort repeated across three temporal windows, and known-state controls comprise 110 jobs and 28.2 billed QPU minutes. Molecular workloads use only 1–5 qubits.
Comparisons with the ideal frozen-state energy separate execution residuals from active-space and ansatz errors. For isolated raw O2 and BeH2 evaluations at 1024 nominal shots, mean residuals are 12–32 times returned standard errors. Pittsburgh BeH2 has cross-job SD 120.4 mHa versus mean returned uncertainty 21.6 mHa. Quadrupling shots approximately halves returned uncertainty without reliably reducing bias. The ZNE package reduces O2 absolute mean residuals by 49.7% on Aachen and 83.5% on Pittsburgh at roughly fourfold charge. Pittsburgh TREX yields greater residual reduction per normalized charge than ZNE; Aachen reverses this ordering. Controls identify qubit-124 preparation/readout discrepancies 5.4–5.6 times calibration estimates. Thirteen mitigated energy estimates fall below mapped spectral minima; no raw estimate does.
The stated limitations include only two devices and three windows, fixed treatment order confounded with time, disabled dynamical decoupling, simplified chemistry models, and unavailable measurement circuits, bitstrings and learned mitigation responses. Controls do not distinguish preparation from assignment error. The shot-allocation model assumes stationary bias and square-root sampling scaling; the proposed diagnostic protocol lacks independent validation.
Variational Quantum Eigensolvers (VQE) promise molecular ground-state energies on today's noisy intermediate-scale quantum (NISQ) processors, but device noise biases the estimated energy severely — under a device-calibrated noise model, the hydrogen molecule's energy error grows from 0.000 mHa (noiseless) to 303.6 mHa (noisy), roughly 190× the "chemical accuracy" threshold of 1.6 mHa that computational chemistry requires. Error-mitigation techniques repair such biases in software, but published results usually report each technique in isolation, making it impossible to know which mitigation to buy with a fixed shot budget. This work answers that question with a controlled comparison. Under identical circuits, noise profiles and shot budgets, I measure the bias and variance of four estimator stacks — raw sampling, twirled-readout extinction (TREX-style calibrated readout inversion), zero-noise extrapolation (ZNE), and ZNE+TREX — across independent random seeds. Three findings emerge. (1) ZNE is a bias–variance tradeoff, not a free lunch: a 7-point Richardson extrapolant reduces the deterministic bias 3200-fold (303.6 → 0.094 mHa) but amplifies statistical noise by a measured factor of up to 25× per run, while TREX removes readout bias with no variance penalty. (2) ZNE has a finite operating window in circuit depth: error recovery falls 99.97% → 99.1% → 95.5% as CX count grows 56 → 112 → 168, and is exhausted entirely for a 1616-CX lithium-hydride circuit at near-term error rates. (3) A noise-bias-tailored repetition code corrects the dominant error axis with logical error rates matching theory (3p², 10p³, 35p⁴ for distances 3, 5, 7) at strictly linear gate overhead (gates = 6d − 6). The results reframe mitigation selection as budget allocation on an empirically measured bias–variance frontier and provide open-source tooling to reproduce every number.
Abstract The inherent noise in current Noisy Intermediate-Scale Quantum (NISQ) devices presents a major obstacle to the accurate implementation of quantum algorithms such as the Variational Quantum Eigensolver (VQE) for quantum chemistry applications. This study examines the impact of error mitigation strategies on VQE performance. We show that, for small molecular systems, an older-generation 5-qubit quantum processing unit (IBMQ Belem), when combined with optimized Twirled Readout Error Extinction (T-REx), achieves ground-state energy estimations an order of magnitude more accurate than those obtained from a more advanced 156-qubit device (IBM Fez) without error mitigation. Our findings demonstrate that T-REx, a computationally inexpensive error mitigation technique, substantially improves VQE accuracy not only in energy estimation, but more importantly in optimizing the variational parameters that characterize the molecular ground state. Consequently, state-vector simulated energies suggest that the accuracy of the optimized variational parameters provides a more reliable benchmark of VQE performance than quantum hardware energy estimates alone. Our results point to the critical role of error mitigation in extending the utility of noisy quantum hardware for molecular simulations.
Quantum error mitigation (QEM) is essential for extracting chemically accurate results from near-term quantum hardware. Many widely used QEM methods rely on uncontrolled heuristics whose bias depends on the specific circuit and noise realization. In this work, we employ QESEM---a characterization-based, unbiased quasi-probabilistic mitigation method---on IBM's Aachen quantum processor to compute the ground-state potential energy surface (PES) of the symmetrically-stretched water molecule. We consider a classically-optimized, single-layer perfect-pairing tiled unitary product state ansatz. We map this ansatz to an 8-qubit register corresponding to a (4,4) active space and the STO-3G basis set. Compared to the statevector reference, we find that raw QPU results typically overestimate the ground-state energy by around 500~mHa across the scanned geometries. On the other hand, QESEM-mitigated results fall approximately within 100~mHa, 30~mHa, or ``chemical accuracy'' ($\sim$1.5~mHa), depending on the target precision. We benchmark QESEM at both loose (0.1~Ha) and tight (0.01~Ha) precision targets, evaluating both individual and merged batches of runs. As the precision target is tightened, the accuracy improves systematically, matching or exceeding results reported in the literature. We further quantify the sampling cost of these results, reporting the number of shots required at each precision level. Our results show how, with the current levels of hardware error, reaching the highest accuracies demands substantial QPU time. The results demonstrate that the characterization-based, unbiased error mitigation provided by QESEM allows to measure quantitatively meaningful potential energy surfaces on current quantum hardware. Concurrently, they highlight the sampling overhead associated with QEM, which remains a central bottleneck en route to larger chemical problems and higher precision.
This study investigates the Variational Quantum Eigensolver (VQE), a hybrid quantum-classical algorithm that partitions the computational workload between a quantum processing unit and a classical optimiser to determine the ground-state energies of molecular systems. The Qiskit SDK v1.x framework is employed, utilising the PySCFDriver module within the qiskit-nature package to construct second-quantised molecular electronic Hamiltonians for two representative systems: Hydrogen ( ) and Lithium Hydride ( ). For the complete active space is retained without spatial reduction, mapping the full Hamiltonian onto a 4-qubit register. However for the full STO-3G representation demands 12 qubits, rendering direct classical parameter optimisation computationally intractable. This bottleneck is resolved by applying an ActiveSpaceTransformer (num_electrons=2, num_spatial_orbitals=3) to freeze the chemically inactive lithium core electrons, reducing the active register to 6 qubits. Fermionic operators are subsequently converted to Pauli strings via the Jordan-Wigner transformation ( ) enabling the quantum processor to evaluate the molecular Hamiltonian as a circuit while preserving the fermionic anti-commutation relations. Under ideal, noise-free simulation using the StatevectorEstimator primitive paired with the COBYLA optimiser (maxiter=200), all VQE energy points converge precisely onto the Full Configuration Interaction (FCI) classical baseline across the full range of interatomic separations. This confirms that the hardware-efficient TwoLocal ansatz ( rotation layers and entanglement layers with reps=3) possesses sufficient expressive power to capture the molecular ground-state wavefunction. Introducing gate-level noise through AerSimulator backends calibrated from the FakeManilaV2 and FakeNairobiV2 hardware profiles—with optimisation driven by the SPSA algorithm (maxiter=150)—reveals two critical failure modes. A systematic upward energy bias is observed, attributable to depolarising gate errors and measurement readout faults that prevent the optimiser from reaching the true minimum energy eigenvalue. Additionally, severe stochastic oscillations between 1.0 Å and 1.5 Å on the curve induce premature convergence of the SPSA optimiser to spurious local minima. These results demonstrate that unmitigated hardware noise renders VQE incapable of reliably predicting equilibrium bond lengths on current NISQ devices, motivating the future integration of Zero-Noise Extrapolation (ZNE), readout error mitigation, and Virtual Distillation into the computational pipeline. Keywords—Variational Quantum Eigensolver (VQE); Qiskit 1.x; Potential Energy Curve; Molecular Ground-State Energy; Hardware-Efficient Ansatz; Gate-Level Noise Simulation; ActiveSpaceTransformer; NISQ Devices.
Quantum processors have crossed the one-hundred-qubit mark, but noise continues to limit circuit performance, while full quantum error correction remains too costly for routine use. Error suppression and mitigation therefore play an important role in extracting value from current hardware, yet independent comparisons of commercial solutions on identical workloads and devices remain scarce. We benchmark IBM Qiskit Runtime, Q-CTRL Performance Management, and Qedma QESEM on IBM Pittsburgh, a 156-qubit IBM Quantum Heron r3 processor. For Sampler workloads, we run Bernstein-Vazirani, quantum phase estimation, GHZ-state preparation, and randomized mirror circuits with up to 100 measured qubits, comparing raw execution, IBM measurement twirling, and Q-CTRL. For Estimator workloads, we measure chain-averaged magnetization and correlation observables of an eight-layer transverse-field Ising circuit at 25, 50, and 75 qubits against an exact matrix-product-state reference, comparing IBM raw execution, IBM TREX plus twirling, Q-CTRL, and QESEM. Q-CTRL produced the best results on the three structured Sampler workloads while keeping reported QPU times within the same order as the IBM configurations. Across six Ising observable and system-size cases, aggregate mean absolute error was 0.0883 for IBM raw execution, 0.0807 for IBM TREX plus twirling, 0.0285 for Q-CTRL, and 0.0188 for QESEM. Relative to raw execution, Q-CTRL and QESEM reduced aggregate error by factors of 3.10 and 4.70, respectively, while QESEM used 7.5 to 11.1 times the reported QPU time of Q-CTRL. These results show that managed error suppression and mitigation can substantially improve current hardware performance, but with distinct accuracy and execution-time tradeoffs.
molecule, emphasizing practical implementation and performance on current quantum hardware. Our research presents a comparative study of HEA and UCCSD ansätze on noiseless and noisy simulations and implements VQE on recent IBM quantum computer noise models and a real quantum computer, IBM Fez, providing a fully functional code employing Qiskit 1.2. Our experiments confirm UCCSD's reliability in ideal conditions, while the HEA demonstrates greater robustness to hardware noise, achieving chemical accuracy on state-vector simulation (SVS). The results reveal that achieving ground-state energy within chemical accuracy is feasible without error mitigation during VQE convergence. We demonstrate that current quantum devices effectively optimize circuit parameters despite misestimating simulated energies. The SVS-evaluated energies provide a more accurate representation of the solution quality compared to QPU-estimated energy values, indicating that VQE converges to the correct ground state despite quantum noise. Our study also applies noise mitigation as a postprocessing technique, using zero-noise extrapolation (ZNE) on a real quantum computer. The detailed methodologies presented in this study, including Hamiltonian construction and Fermionic-to-qubit transformations, facilitate flexible adaptation of the VQE approach for various algorithm variants and across different levels of algorithmic implementation.