Hamiltonian-reconstruction distance as a success metric for the Variational Quantum Eigensolver

Leo Joon Il Moon, Mandar M. Sohoni, Michael A. Shimizu, Praveen Viswanathan, Kevin Zhang, Eun-Ah Kim, Peter L. McMahonarXiv 2024 · DOI · arXiv

Abstract

The Variational Quantum Eigensolver (VQE) is a hybrid quantum-classical algorithm for quantum simulation that can be run on near-term quantum hardware. A challenge in VQE—as well as any other heuristic algorithm for finding ground states of Hamiltonians—is to know how close the algorithm's output solution is to the true ground state when the true ground state and ground-state energy are unknown. This is especially important in iterative algorithms, such as VQE, where one wants to avoid erroneous early termination. Recent developments in Hamiltonian reconstruction—the inference of a Hamiltonian given an eigenstate—give a metric that can be used to assess the quality of a variational solution to a Hamiltonian-eigensolving problem. This metric can assess the proximity of the variational solution to the ground state without any knowledge of the true ground state or ground-state energy. In numerical simulations and in demonstrations on a cloud-based trapped-ion quantum computer, we show that for examples of both one-dimensional transverse-field-Ising (11 qubits) and two-dimensional J1-J2 transverse-field-Ising (6 qubits) spin problems, the Hamiltonian-reconstruction distance gives a helpful indication of whether VQE has yet found the ground state or not. Our experiments included cases where the energy plateaus as a function of the VQE iteration, which could have resulted in erroneous early stopping of the VQE algorithm, but where the Hamiltonian-reconstruction distance correctly suggests to continue iterating. We find that the Hamiltonian-reconstruction distance has a useful correlation with the fidelity between the VQE solution and the true ground state. Our work suggests that the Hamiltonian-reconstruction distance may be a useful tool for assessing success in VQE, including on noisy quantum processors in practice.

Introduction

The Variational Quantum Eigensolver (VQE) is a primary algorithm for the Noisy Intermediate-Scale Quantum (NISQ) era, used to find the ground state of complex molecules and materials. However, VQE suffers from a significant diagnostic challenge: because it is an iterative heuristic, it is difficult to determine when the algorithm has truly converged to the global minimum. In practice, researchers often terminate the process when the energy reaches a plateau. However, in complex energy landscapes or under the influence of quantum noise, such plateaus may represent local minima or "barren plateaus" rather than the true ground state. Without prior knowledge of the ground state energy, a reliable "yardstick" is needed to verify the quality of the solution.

Methodology: Hamiltonian Reconstruction

To address this "blind spot," we propose the use of Hamiltonian Reconstruction (HR) distance ($\Delta$) as a success metric. This approach effectively reverse-engineers the physics: given a quantum state produced by VQE, it identifies the Hamiltonian for which that state would be an eigenstate.

Overview of the VQE process integrated with Hamiltonian Reconstruction (HR) to compute the distance metric Δ.

As shown in the workflow above, the process involves measuring a set of operators and their correlators to construct a covariance matrix. The reconstructed Hamiltonian ($H_R$) is derived from this matrix, and the HR distance ($\Delta$) is calculated as the Euclidean (L2) distance between the coefficients of the target Hamiltonian and the reconstructed one. If the VQE state is a true eigenstate of the target Hamiltonian, $\Delta$ should be zero. Crucially, this metric relies only on experimental measurements and the known target Hamiltonian, requiring no prior knowledge of the ground state energy.

Experimental and Numerical Validation

The metric was evaluated using two spin-model benchmarks: an 11-qubit 1D Transverse-Field Ising Model (TFIM) and a 6-qubit 2D $J_1-J_2$ TFIM. Experiments were conducted both in numerical simulations and on IonQ’s "Harmony" trapped-ion quantum computer.

VQE performance metrics (Energy, Fidelity, and HR distance) for 1D and 2D spin models in simulation and on hardware.

Key Findings

  • Detecting False Convergences: The HR distance successfully flags suboptimal solutions even when energy appears stable. In 11-qubit simulations (Fig. 2a), the energy plateaued around iteration 600, but the HR distance remained high (~0.4), correctly signaling that the state was not yet the ground state. This prevents erroneous early termination.
  • Correlation with Fidelity: There is a strong negative correlation between HR distance and state fidelity (the overlap with the true ground state). As fidelity increases, the HR distance consistently decreases.
  • Hardware Performance: On the IonQ Harmony processor, the HR distance successfully tracked solution quality for both models. While robust, the metric requires a sufficient measurement budget (approximately $10^4$ shots) to maintain a reliable signal-to-noise ratio.
  • Noise Sensitivity: Simulations indicate the metric remains useful until gate error probabilities exceed approximately 10%, at which point the correlation between HR distance and state fidelity begins to degrade.

Conclusion

The Hamiltonian-reconstruction distance provides a bounded (0 to 1), interpretable metric for assessing VQE success on NISQ hardware. Unlike traditional methods, it does not require knowing the answer in advance, making it a powerful diagnostic tool for real-world quantum simulations. Future work may explore using HR distance directly within the VQE cost function to help optimizers navigate complex energy landscapes more effectively.

Summary updated

Abstract

The Variational Quantum Eigensolver (VQE) is a hybrid quantum-classical algorithm for quantum simulation that can be run on near-term quantum hardware. A challenge in VQE -- as well as any other heuristic algorithm for finding ground states of Hamiltonians -- is to know how close the algorithm's output solution is to the true ground state, when the true ground state and ground-state energy are unknown. This is especially important in iterative algorithms, such as VQE, where one wants to avoid erroneous early termination. Recent developments in Hamiltonian reconstruction -- the inference of a Hamiltonian given an eigenstate -- give a metric can be used to assess the quality of a variational solution to a Hamiltonian-eigensolving problem. This metric can assess the proximity of the variational solution to the ground state without any knowledge of the true ground state or ground-state energy. In numerical simulations and in demonstrations on a cloud-based trapped-ion quantum computer, we show that for examples of both one-dimensional transverse-field-Ising (11 qubits) and two-dimensional J1-J2 transverse-field-Ising (6 qubits) spin problems, the Hamiltonian-reconstruction distance gives a helpful indication of whether VQE has yet found the ground state or not. Our experiments included cases where the energy plateaus as a function of the VQE iteration, which could have resulted in erroneous early stopping of the VQE algorithm, but where the Hamiltonian-reconstruction distance correctly suggests to continue iterating. We find that the Hamiltonian-reconstruction distance has a useful correlation with the fidelity between the VQE solution and the true ground state. Our work suggests that the Hamiltonian-reconstruction distance may be a useful tool for assessing success in VQE, including on noisy quantum processors in practice.
BibTeX
@misc{moon2024hamiltonian,
  author = {Leo Joon Il Moon and Mandar M. Sohoni and Michael A. Shimizu and Praveen Viswanathan and Kevin Zhang and Eun-Ah Kim and Peter L. McMahon},
  title = {{Hamiltonian-reconstruction distance as a success metric for the Variational Quantum Eigensolver}},
  year = {2024},
  eprint = {2403.11995},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2403.11995},
}
Figure