Quantum resource reduction for quantum-centric supercomputing via correlated mean-field downfolding framework
Thien Ngoc Tran, Lan Nguyen Tran · 2026
Read Quantum resource reduction for quantum-centric supercomputing via correlated mean-field downfolding framework by Thien Ngoc Tran, Lan Nguyen Tran (2026) 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-2605-08675.txt.
Introduction
Thien Ngoc Tran and Lan Nguyen Tran cut the quantum cost of working out how a molecule acts. A first step on a normal computer folds the effect of outer electron paths into a one body term. That term keeps the same form as the bare one. So the quantum part needs no extra circuits. A sampling step then solves the folded form. In this work the sampling ran on a simulator, not real quantum hardware.
Tests used a chain, ring, and grid of hydrogen atoms. Tests also used the nitrogen molecule. At the same active size, the folded form beat the plain form each time. The trick is simple to add. It should stretch to solids as well.
Catalog record
| CREATOR | Thien Ngoc Tran, Lan Nguyen Tran |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 26 KB |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2605.08675
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Quantum resource reduction for quantum-centric supercomputing via correlated mean-field downfolding framework Thien Ngoc Tran, Lan Nguyen Tran Abstract We present OBDF-SQD, a hybrid quantum-classical method that combines one-body downfolding (OBDF) based on one-body Møller–Plesset second-order perturbation theory (OBMP2) with sample-based quantum diagonalization (SQD) for use in quantum-centric supercomputing (QCS). In this approach, OBMP2 is executed classically to fold dynamical correlation from external orbitals into a renormalized one-body operator, yielding an effective active-space Hamiltonian that retains the same operator structure as the bare Hamiltonian and therefore requires no additional quantum circuit resources. SQD is then applied to this effective Hamiltonian, where, in this work, the quantum sampling is performed via the Qiskit Aer simulator rather than actual quantum hardware. We benchmark OBDF-SQD on dissociation curves of H 6 \text{H}{\vphantom{\text{X}}}_{\vphantom{\text{2}}\smash[t]{\text{6}}}^{\vphantom{\smash[t]{\text{2}}}\hphantom{\text{6}}\text{}} chain, ring, and lattice systems and the N 2 \text{N}{\vphantom{\text{X}}}_{\vphantom{\text{2}}\smash[t]{\text{2}}}^{\vphantom{\smash[t]{\text{2}}}\hphantom{\text{2}}\text{}} molecule in the cc-pVDZ basis, comparing against standard methods and active-space SQD (CAS-SQD). We observed that OBDF-SQD consistently improves upon CAS-SQD with the same active space. The simplicity of the one-body downfolding correction also makes the approach straightforwardly extensible to periodic solids within existing quantum embedding frameworks. I Introduction Quantum computing is highly promising for simulating complicated molecules and materials challenging to the classical methods 5 ; 30 ; 9 ; 34 . A central vision guiding current development is that of quantum-centric supercomputing (QCS) 8 ; 22 , in which quantum processors are tightly integrated with classical high-performance computing (HPC) resources. In this framework, quantum hardware serves as a specialized co-processor that handles computationally intractable tasks, while classical supercomputers manage the remaining workload. This framework is particularly well-suited to near-term noisy intermediate-scale quantum (NISQ) devices, which remain limited by noise and decoherence 13 , and motivates the development of hybrid quantum-classical algorithms that are robust to hardware imperfections and can scale alongside improvements in both quantum and classical computing. A prominent family of hybrid quantum-classical algorithms is built around the variational quantum eigensolver (VQE) 39 ; 31 ; 20 , in which a parameterized quantum circuit prepares a trial state and the expectation value of the Hamiltonian is measured on the quantum device, while a classical optimizer updates the circuit parameters. The unitary coupled-cluster (UCC) ansatz 4 ; 48 ; 41 ; 1 has been widely adopted as a chemically motivated state-preparation ansatz for VQE. Substantial effort has been devoted to reducing the quantum resources required by VQE, such as qubit coupled-cluster (QCC) 43 ; 42 , ADAPT-VQE 17 ; 47 ; 58 , and factorized UCC ansatz 11 ; 55 . Other adaptive schemes inspired by classical quantum chemistry have also been proposed 15 ; 45 . Complementary strategies reduce the problem size by partitioning the system into active spaces or subsystems. Nakagawa and co-workers proposed deep VQE 16 ; 33 , dividing the system into subsystems each solved independently with VQE. Quantum embedding frameworks, including density matrix embedding theory (DMET) 32 ; 49 ; 23 , dynamical mean-field theory (DMFT) 21 ; 6 , and density functional embedding 14 ; 18 , have further extended the reach of VQE to larger systems. Despite these advances, VQE faces persistent practical difficulties: the classical optimization loop is costly, convergence can be hindered by barren plateaus, and circuit depths required for chemically accurate results remain challenging on current hardware. Sample-based quantum diagonalization (SQD) 40 ; 3 ; 12 ; 19 ; 44 offers an alternative route to hybrid quantum-classical computation that resolves many of these difficulties. Rather than optimizing a parameterized circuit, SQD leverages repeated projective measurements, bitstring samples, obtained from a quantum device to construct a compressed configuration-interaction (CI) subspace, which is then diagonalized classically to approximate the ground-state energy. Because SQD requires no classical optimization of circuit parameters and naturally delegates the diagonalization step to classical HPC resources, it aligns well with the QCS model. SQD has been demonstrated on real quantum hardware to achieve chemically accurate results for molecules such as the nitrogen dimer and iron-sulfur clusters 40 , systems that pose a formidable challenge to classical methods. A key limitation shared by both VQE and SQD in active-space formulations is the incomplete treatment of dynamical electron correlation outside the active space. Neglecting this contribution leads to systematic errors in energies, particularly for small active spaces. To address this challenge, Kowalski and co-workers employed downfolding based on the double UCC (DUCC) ansatz to construct effective active-space Hamiltonians 7 ; 24 that integrate out high-energy degrees of freedom while reproducing exact energies. In the same vein, we have recently developed the one-body downfolding (OBDF) in a resource-efficient way 27 . In the OBDF framework based on one-body Møller-Plesset second-order perturbation theory (OBMP2) 50 ; 51 ; 52 ; 53 ; 54 , the effect of external (inactive) orbitals is incorporated into the active-space Hamiltonian through a correlated one-electron operator that modifies only the one-body part of the Hamiltonian. The resulting effective Hamiltonian retains the same operator structure as the bare active-space Hamiltonian and therefore requires no additional quantum resources beyond those needed for the bare problem. OBDF can be derived from the DUCC ansatz by restricting the external cluster operator to double excitations involving at least one inactive index and truncating the Baker-Campbell-Hausdorff (BCH) expansion at second order. This makes OBDF a natural and practical component within the QCS framework: the additional computational cost is entirely classical, while the quantum circuit complexity is unchanged. In this work, we combine SQD with the OBDF framework (OBDF-SQD) to simultaneously address noise tolerance, active-space accuracy, and dynamical correlation. The method proceeds as follows: OBMP2 is first run classically for the full system to generate correlated molecular orbitals and to construct the effective active-space Hamiltonian encoding dynamical correlation outside the active space through a modified one-body term; SQD is then applied to this effective Hamiltonian, with the quantum sampling performed through the Qiskit Aer simulator in place of real quantum hardware. We benchmark OBDF-SQD on the dissociation of H 6 chain, ring, and lattice in the cc-pVDZ basis set, as well as the N 2 molecule, comparing against FCI, CCSD(T), HF, OBMP2, and standard CAS-SQD. Our results demonstrate that OBDF-SQD systematically improves upon CAS-SQD with the same active space. We also discuss the current limitations of the OBDF framework and highlight promising directions for future work. II Theory II.1 Sample-Based Quantum Diagonalization The many-body electronic Hamiltonian in second quantization reads \hat{H}=\sum_{pq\sigma}h_{pq}\,\hat{a}^{\dagger}_{p\sigma}\hat{a}_{q\sigma}+\frac{1}{2}\sum_{pqrs\,\sigma\eta}g_{pq}^{rs}\,\hat{a}^{\dagger}_{p\sigma}\hat{a}^{\dagger}_{r\eta}\hat{a}_{q\sigma}\hat{a}_{s\eta}, (1) where h_{pq} and g_{pq}^{rs} are the one- and two-electron integrals, respectively. The Jordan–Wigner (JW) transformation maps the fermionic operators onto Pauli strings acting on 2N_{\mathrm{MO}} qubits, with computational basis states corresponding to Slater determinants encoded as bitstrings |x\rangle=(x_{\downarrow},x_{\uparrow}) , where x_{p\sigma}\in\{0,1\} denotes the occupation of the p -th spin orbital. In the SQD framework 40 ; 3 ; 12 ; 19 ; 44 , quantum circuits employing the Local Unitary Cluster Jastrow (LUCJ) ansatz 35 are used to generate a set of sampled bitstring configurations \tilde{\mathcal{X}} . In this work, the quantum sampling is performed via the Qiskit Aer simulator as an emulator of physical quantum hardware 2 . Particle-number conservation, N_{x}=N , is enforced to filter the valid subset \mathcal{X}_{N}\subset\tilde{\mathcal{X}} . Configurations violating this constraint are subjected to a self-consistent recovery procedure, wherein bits are probabilistically flipped according to a modified