Hybrid Quantum-Classical Eigensolver with Real-Space Sampling and Symmetric Subspace Measurements
Lei Xu, Ling Wang · 2025
Read Hybrid Quantum-Classical Eigensolver with Real-Space Sampling and Symmetric Subspace Measurements by Lei Xu, Ling Wang (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-2510-19219.txt.
Introduction
Lei Xu and Ling Wang propose a symmetry-adapted solver for strongly correlated quantum systems that separates where a wave function is sampled from where it is defined. Parallel quantum circuits handle entanglement inside blocks of a lattice, while a matrix product backbone links the blocks and restores global correlations.
Configurations are drawn directly from the sampled amplitudes, block by block, with no Markov chain needed. The variational state itself lives in a symmetry-adapted basis, where each class of equivalent configurations gathers its total weight. Mixing quantum circuits, tensor networks, and symmetry in this way targets accurate energies for hard materials.
Catalog record
| CREATOR | Lei Xu, Ling Wang |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 29 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2510-19219
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Hybrid Quantum-Classical Eigensolver with Real-Space Sampling and Symmetric Subspace Measurements Lei Xu, Ling Wang Abstract We propose a hybrid quantum-classical eigensolver to address the computational challenges of simulating strongly correlated quantum many-body systems, where the exponential growth of the Hilbert space and extensive entanglement render classical methods intractable. Our approach combines real-space sampling of tensor-network-bridged quantum circuits with symmetric subspace measurements, effectively constraining the wavefunction within a substaintially reduced Hilbert space for efficient and scalable simulations of versatile target states. The system is partitioned into equal-sized subsystems, where quantum circuits capture local entanglement and tensor networks reconnect them to recover global correlations, thereby overcoming partition-induced limitations. Symmetric subspace measurements exploit point-group symmetries through a many-to-one mapping that aggregates equivalent real-space configurations into a single symmetric state, effectively enhancing real-space bipartition entanglement while elimilating redundant degrees of freedom. The tensor network further extends this connectivity across circuits, restoring global entanglement and correlation, while simultaneously enabling generative sampling for efficient optimization. As a proof of concept, we apply the method to the periodic J 1 − J 2 J_{1}\!-\!J_{2} antiferromagnetic Heisenberg model in one and two dimensions, incorporating translation, reflection, and inversion symmetries. With a small matrix product state bond dimension of up to 6, the method achieves an absolute energy error of 10 − 5 10^{-5} for a 64-site periodic chain and a 6 × 6 6\times 6 torus after bond-dimension extrapolation. These results validate the accuracy and efficiency of the hybrid eigensolver and demonstrate its strong potential for scalable quantum simulations of strongly correlated systems. 1 Introduction Simulating strongly correlated quantum many-body systems on classical computers remains a formidable challenge due to the exponential growth of the Hilbert space with system size. This vast dimensionality renders brute-force methods such as exact diagonalization (ED) infeasible for large systems SandvikAW2010_lecture ; VieijraT2020 ; WeisseA2008a ; LaflorencieN2004a ; LauchliAM2011 . Tensor network approaches, including matrix product states (MPS), offer an efficient alternative by compressing the wavefunction PhysRevLett.93.040502 ; PhysRevLett.99.220405 ; VerstraeteCirac2004PEPS ; PhysRevLett.93.227205 ; VerstraeteMurgCirac2008AdvPhys ; however, accurately capturing entanglement beyond one-dimensional systems requires rapidly increasing bond dimensions. The resulting memory and computational costs quickly become prohibitive, underscoring the need for novel strategies to efficiently represent quantum correlations. Quantum computation offers a promising route to alleviate the classical bottlenecks in simulating quantum systems BiamonteJ2017 ; PreskillJ2018 ; CerezoM2021 . A particularly appealing approach is to partition large quantum circuits into smaller subcircuits that can be executed in parallel across multiple quantum processing units. In such a distributed architecture, quantum and classical resources operate synergistically: quantum subcircuits capture local correlations, while classical tensor-network coordination integrates their outputs to reconstruct the global quantum state BravyiS2016 ; EddinsA2022 ; NiuJ2023 ; CarreraVazquezA2024 ; EndoS2021 . Although this collective quantum-classical workflow effectively addresses the scaling challenge associated with qubit numbers, its ability to faithfully represent strongly correlated systems with substantial entanglement remains limited. Nevertheless, it provides a viable and efficient framework for moderately correlated or weakly entangled problems. To overcome these entanglement limitations, we introduce a hybrid quantum-classical framework that integrates real-space sampling with symmetric-subspace measurements. Real-space sampling exploits the generative capability of tensor-backed parallel quantum circuits FerrisAJ2012 ; HanZY2018 , enabling efficient exploration of spatial configurations. Meanwhile, symmetric-subspace measurements project the quantum state onto a drastically reduced symmetric Hilbert space, effectively filtering out irrelevant degrees of freedom while establishing long-range entanglement between distant real-space components ChooK2018 ; ChooK2019 ; BaoST2025 ; Bukov_21 . This projection simultaneously compresses the relevant Hilbert space and enriches the representation of real-space correlations, allowing the hybrid framework to capture strongly correlated many-body features with high efficiency. The core idea behind this hybrid approach is a many-to-one mapping between real-space and symmetric bases, defined as \widetilde{p}(a_{\text{symm}})=\sum_{g\in G}p(a_{\text{real}}\,|\,a_{\text{repr}}=ga_{\text{real}}) , where p(a_{\text{real}}) and \widetilde{p}(a_{\text{symm}}) denote the probability distributions in the real and symmetric bases, respectively, g is a group element in symmetry group G , and there exists a one-to-one correspondence between a_{\text{symm}}\leftrightarrow a_{\text{repr}} . Intuitively, this mapping aggregates the probabilities of all symmetry-equivalent real-space configurations, assigning their collective contribution to the corresponding symmetric basis state a_{\text{symm}} . The summation coherently removes phase interference among configurations related by symmetry operations, while retaining relative phases between different a_{\text{repr}} as encoded by the tensor-backed quantum circuits. As a result, measurements performed in the symmetric basis \{a_{\text{symm}}\} substantially reduce the entanglement burden on the tensor network. In addition, the generative nature of real-space sampling obviates the need for Markov chain updates, yielding significant improvements in both computational efficiency and scalability. To validate the effectiveness of our hybrid quantum-classical framework, we apply it to the periodic J_{1}\!-\!J_{2} antiferromagnetic Heisenberg model in one and two dimensions—a benchmark system for studying frustration and quantum correlations. In our demonstration, translation, reflection, inversion, and particle number conservation symmetries are incorporated to ensure that both sampling and measurement procedures fully respect the system’s inherent symmetries. With a small matrix product state bond dimension of up to 6, the method achieves an absolute energy error of 10^{-5} for a 64-site periodic chain and a 6\times 6 torus after bond-dimension extrapolation. These results confirm the accuracy and consistency of the approach, while indicating strong potential for scaling to larger and more complex quantum systems in future studies. 2 Method 2.1 Wavefunction We partition an N -site system into n_{b} blocks, each containing b spins, such that N=bn_{b} . A real-space configuration is denoted as a_{\rm real}=\otimes_{i=1}^{n_{b}}\bar{a}_{i} , where each block state \bar{a}_{i}=\otimes_{k=(i-1)b}^{ib-1}s_{k} . The corresponding real-space distribution function p(a_{\rm real})\equiv|\phi(a_{\rm real})|^{2} is generated by two coupled components: a set of parallel quantum circuits {C_{i}} and a MPS backbone {B_{i}} . Each circuit C_{i} projects the local spin block \bar{a}_{i} onto a renormalized basis \gamma_{i} , while the MPS tensors {B_{i}} integrate these renormalized bases {\gamma_{1},\ldots,\gamma_{n_{b}}} to produce the final wave-function amplitude \phi(a_{\rm real})=\sum_{\{\mu_{i},\gamma_{i}\}}B_{\mu_{1}}^{\gamma_{1}}C_{\gamma_{1}}[\bar{a}_{1}]B_{\mu_{1}\mu_{2}}^{\gamma_{2}}C_{\gamma_{2}}[\bar{a}_{2}]\cdots B_{\mu_{n_{b}-1}}^{\gamma_{n_{b}}}C_{\gamma_{n_{b}}}[\bar{a}_{n_{b}}]. (1) This hybrid structure is illustrated in Fig. 1 (a), where each C_{i} represents a local unitary transformation satisfying C_{i}^{\dagger}C_{i}=I_{\chi\times\chi} , with \chi\leq 2^{b} the number of retained renormalized states. Note that \phi(a_{\rm real}) does not yet represent the final physical wave function, as subsequent measurements are performed within the symmetric subspace. To define the target symmetric subspace and corresponding wave function, we first construct an orthogonal basis from symmetry-constrained superpositions of the Ising configurations |a_{\rm real}\rangle=|s_{1}\otimes s_{2}\otimes\cdots\otimes s_{N}\rangle . Taking one-dimensional translational symmetry as an example, a symmetric basis vector is defined as BaoST2025 ; SandvikAW2010_lecture ; VieijraT2020 ; WeisseA2008a ; LaflorencieN2004a ; LauchliAM2011 |a_{\rm symm}\rangle=\frac{1}{\sqrt{N_{a_{\rm repr}}}}\sum_{r=0}^{N-1}e^{-ikr}T^{r}|a_{\rm repr}\rangle, (2) where T denotes the lattice translation operator and N_{a_{\rm repr}} is a normalization factor, such that \langle a_{\rm symm}|a_{\rm symm}\rangle=1 . Each representative configuration |a_{\rm repr}\rangle is the lexicographically smallest element within the equivalence class \{T^{r}|a_{\rm repr}\rangle\} , where spins are encoded in binary form ( 0 for |\uparrow\rangle , 1 for |\downarrow\rangle ). For a given momentum sector k , a state |\psi\rangle satisfying T|\psi\rangle=e^{ik}|\psi\rangle can be expanded as |\psi\rangle=\sum_{a_{\rm symm}=1}^{N_{\text{repr}}}\psi(a_{\rm symm})|a_{\rm symm}\rangle, (3) where N_{\text{repr}} is the dimension of the symmetric subspace, which is substantially smaller than the full Hilbert space due to the imposed symmetry and orthogonality constraints. This construction can be straightforwardly generalized to higher dimensions and other symmetry groups. Our goal is to transform the real-space distribution function \phi(a_{\rm real}) defined in Eq.( 1 ) into