Excitation Amplitude Sampling for Low Variance Electronic Structure on Quantum Computers
Connor Lenihan 1, Oliver J. Backhouse 1, Tom W. A. Montgomery 2, Phalgun Lolur 2, M. J. Bhaseen 1, George H. Booth 1 · 2025
Read Excitation Amplitude Sampling for Low Variance Electronic Structure on Quantum Computers by Connor Lenihan 1, Oliver J. Backhouse 1, Tom W. A. Montgomery 2, Phalgun Lolur 2, M. J. Bhaseen 1, George H. Booth 1 (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2506-15438.txt.
Introduction
Researchers propose in 2025 a cheaper way to read energy from quantum states of electrons in molecules. Reading a full state takes vast numbers of repeats, called shots. Their method blends classic rules of thumb with partial scans of the state.
A fixed set of excitation strengths is sampled from the state, then fed into an energy formula. Tests show near a hundredfold drop in shots for the same error size. The method also holds up well on real quantum hardware, bearing far more noise than older reads. Gains grow in a straight line with system size in the range tested. A linked classic model can also pull out traits beyond energy alone.
Catalog record
| CREATOR | Connor Lenihan 1, Oliver J. Backhouse 1, Tom W. A. Montgomery 2, Phalgun Lolur 2, M. J. Bhaseen 1, George H. Booth 1 |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 92 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2506-15438
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Excitation Amplitude Sampling for Low Variance Electronic Structure on Quantum Computers Connor Lenihan 1, Oliver J. Backhouse 1, Tom W. A. Montgomery 2, Phalgun Lolur 2, M. J. Bhaseen 1, George H. Booth 1 Abstract We combine classical heuristics with partial shadow tomography to enable efficient protocols for extracting information from correlated ab initio electronic systems encoded on quantum devices. By proposing the use of a correlation energy functional and sampling of a polynomial set of excitation amplitudes of the quantum state, we can demonstrate an almost two order of magnitude reduction in required number of shots for a given statistical error in the energy estimate, as well as observing a linear scaling to accessible system sizes. Furthermore, we find a high-degree of noise resilience of these estimators on real quantum devices, with up to an order of magnitude increase in the tolerated noise compared to traditional techniques. While these approaches are expected to break down asymptotically, we find strong evidence that these large system arguments do not prevent algorithmic advantage from these simple protocols in many systems of interest. We further extend this to consider the extraction of beyond-energetic properties by mapping to a coupled cluster surrogate model, as well as a natural combination within a quantum embedding framework. This embedding framework avoids the unstable self-consistent requirements of previous approaches, enabling application of quantum solvers to realistic correlated materials science, where we demonstrate the volume-dependence of the spin gap of Nickel Oxide. I Introduction The simulation of physical quantum systems is one of the key applications driving the development of digital quantum devices, and an original motivation for their inception Feynman 1982 ; Bauer et al. 2020 ; Cao et al. 2019 ; McArdle et al. 2020 . While there has been a huge growth of algorithms to leverage the potential for quantum information processing, a significant challenge limiting their application to realistic systems is the extraction of observables. These observables are cast as the expectation values of samples from a probability distribution via measurements of the quantum device, and the overhead and associated variance with these repeated measurements significantly limits the scale and scope of the information which can be extracted from the simulated system. The most important observable is the energy of a quantum state. This can be estimated via a digital readout when quantum phase estimation (QPE) and related algorithms act on a (near) eigenstate of the system Hamiltonian Kitaev 1995 ; Kivlichan et al. 2020 ; Lin and Tong 2022 ; Blunt et al. 2023 , or via a direct sampling of the expectation value. This direct sampling approach relies on the low-body nature of ab initio interactions in a quantum system, to decompose the Hamiltonian into a weighted average of a polynomial (yet large) number of Pauli strings. These can be estimated and averaged over samples before combination to a statistical estimate of the energy. This is a core primitive routine at the heart of current and near-term ‘NISQ’ algorithms, which trade circuit depth and qubits for increased circuit repetitions and measurements (shots) Preskill 2018 ; McClean et al. 2016 ; Tilly et al. 2022 ; Blekos et al. 2024 . However, it was realized early on that for realistic systems, the \mathcal{O}[N^{4}] native Pauli strings of the ab initio Hamiltonian would result in a ‘variance problem’ Wecker et al. 2015 . This is due to the large number of shots required, M , to reach the low statistical uncertainty in the energy estimate required for chemical and materials science applications. Intense research has therefore taken place in recent years to reduce the number of shots for an associated variance of the estimate, based on grouping the Pauli strings into commuting cliques Izmaylov et al. 2019 ; Jena et al. 2019 ; Crawford et al. 2021 ; Zhao et al. 2020 ; Gokhale et al. 2020 ; Yen et al. 2020 ; Verteletskyi et al. 2020 ; Wu et al. 2023 ; Yen et al. 2023 ; Yen and Izmaylov 2021 ; Bonet-Monroig et al. 2020a ; Burns et al. 2024 , using randomized approaches Huang et al. 2020 ; Elben et al. 2023 ; Hadfield et al. 2022 ; Dutt et al. 2023 ; Zhao et al. 2021 ; Bian and Wu 2025 , adaptive strategies Shlosberg et al. 2023 ; Zhu et al. 2024a ; Zhu et al. 2024b ; Gresch et al. 2025 ; Gresch and Kliesch 2025 , factorized operators Huggins et al. 2021 , or leveraging machine learned or signal processing models Liang et al. 2024 ; Iouchtchenko et al. 2023 ; Bassi et al. 2025 . For an additive error in the energy estimate, \epsilon , early forecasts for a minimal-basis 14-qubit water molecule, resulted in an estimate of M\sim 10^{9} for a modest \epsilon=1 mE h statistical error, despite this system being simple to solve exactly on classical resources Wecker et al. 2015 . More recent state of the art variance reduction methods have reduced this estimate to M\sim 10^{8} shots Huggins et al. 2021 ; Mazzola 2024 , or M\sim 10^{7} for the slightly higher target error of 1.6mE h (taken as chemical accuracy) Gresch and Kliesch 2025 . For a more challenging (albeit still relatively small) Ferredoxin target system (Fe 2 O 2 ) in 56 orbitals, an estimate of 10^{13} shots was given per single energy estimate Wecker et al. 2015 . While estimates of long-term circuit execution times vary with both hardware and details of the circuit and error mitigation or correction overheads Sunada et al. 2022 ; Lee et al. 2021 , these required shot numbers and their scaling with system size constitute a severe limitation in the scope of systems that can be considered Bansingh et al. 2022 ; Peng et al. 2023 ; McNulty et al. 2023 . This is often considered as motivation for QPE-like algorithms with a digital readout of the energy. These are expected to come to the fore once fault tolerance of the devices permits the depth of circuits required, which will scale linearly with the inverse precision, in a Heisenberg-limited fashion as \sim 1/\epsilon Bay-Smidt et al. 2025 . This allows for an energy estimate which avoids an explicit dependence on a number of shots (although there will likely still be a dependence from the initial state preparation problem Lee et al. 2023 ). However, this does not entirely solve the problem, since the extraction of other critically important expectation values for characterizing electronic structure, such as electrical moments, reduced density matrices, nuclear forces and other response properties are not able to be extracted in this fashion. Returning to a direct sampling of these associated operators could still be required O’Brien et al. 2022 ; Chen et al. 2021 . As a final motivating challenge for the developments of this work, we consider the importance of coupling quantum resources to the description of a wider system of interest on classical resources, enabling multi-scale and multi-resolution hybrid quantum-classical workflows. For application of even long-term quantum algorithms to realistic chemical and materials science applications, these hybrid approaches are likely to be essential, allowing for the targeted application of precious quantum resources to subspaces of systems featuring particularly