arxiv: 2604.06741 · v2 · submitted 2026-04-08 · ❄️ cond-mat.dis-nn · cond-mat.str-el· physics.chem-ph

Recognition: 2 theorem links

· Lean Theorem

Projector, Neural, and Tensor-Network Representations of mathbb{Z}_N Cluster and Dipolar-cluster SPT States

Daesik Kim, Hyun-Yong Lee, Jung Hoon Han, Seungho Lee

Authors on Pith no claims yet

Pith reviewed 2026-05-10 17:43 UTC · model grok-4.3

classification ❄️ cond-mat.dis-nn cond-mat.str-elphysics.chem-ph

keywords Z_N cluster statessymmetry-protected topological ordertensor product statesneural quantum statesKramers-Wannier dualitymatrix product statesdipolar symmetries

0 comments

The pith

The P-representation efficiently encodes Z_N cluster and dipolar-cluster SPT wavefunctions and yields closed-form neural weights plus a three-index tensor product form.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper establishes that the Z_N cluster-state wavefunction, a standard example of symmetry-protected topological order, admits an efficient projector-based description called the P-representation. This description rewrites the local interaction matrix through a restricted Boltzmann machine construction, producing an exact closed-form weight function W(s,h) that couples physical spins to hidden variables and thereby generates both neural quantum states and matrix product states. For the dipolar-cluster variant protected by charge and dipole symmetries, the same procedure produces a tensor product state in which every local tensor carries three virtual indices linking each site to its two nearest neighbors and one further neighbor. Density-matrix renormalization group benchmarks indicate that this tensor product representation can be more efficient than conventional matrix product states for certain modulated SPT phases. As a side result the authors extend the Kramers-Wannier operator to Z_N and reinterpret it as a non-invertible dipolar generalization of the discrete Fourier transform.

Core claim

The Z_N cluster-state wavefunction is expressed in the projector-based P-representation, identified as the efficient scheme. The interaction matrix in this representation is rewritten via the restricted Boltzmann machine scheme in terms of neural weight matrices to obtain neural quantum state and matrix product state forms of the identical state. Closed-form weight functions W(s,h) coupling physical spins s to hidden variables h are derived for both the ordinary Z_N cluster state and the dipolar-cluster state. The dipolar case directly produces a tensor product state representation in which each local tensor carries three virtual indices. The resulting tensor product construction is compared

What carries the argument

The P-representation, a projector-based scheme that encodes the cluster wavefunction through local constraints between physical spins s and auxiliary hidden variables h, which is then recast into neural weights W(s,h) and grouped into tensor networks.

If this is right

The same weight matrices can be grouped either as a neural network or as a matrix product state, giving two numerically distinct but mathematically equivalent representations of one wavefunction.
The tensor product state for the dipolar cluster state connects each site to two nearest neighbors plus one further neighbor through three virtual indices.
The generalized Z_N Kramers-Wannier operator is non-invertible because it functions as a dipolar discrete Fourier transform on the Z_N variables.
Density-matrix renormalization group comparisons suggest the tensor product state can be computationally lighter than standard matrix product states for some modulated SPT phases.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

The same projector-to-weight-function route may supply explicit tensor representations for other symmetry-protected phases that involve dipole or higher multipole symmetries.
The three-index local tensors could reduce the effective bond dimension needed along certain directions when the SPT pattern is spatially modulated.
Because the construction is exact and parameter-free, it supplies a controlled testbed for studying how neural-network and tensor-network ansatzes capture topological entanglement.

Load-bearing premise

Rewriting the interaction matrix of the P-representation through the restricted Boltzmann machine scheme produces an exact representation that preserves the topological properties without introducing approximations or parameter adjustments that would alter the symmetries.

What would settle it

Exact computation of the overlap between the constructed TPS wavefunction and the known analytic cluster or dipolar-cluster state on a finite open chain; any systematic drop below unity as system size or N increases would show the representation is not faithful.

Figures

Figures reproduced from arXiv: 2604.06741 by Daesik Kim, Hyun-Yong Lee, Jung Hoon Han, Seungho Lee.

**Figure 3.** Figure 3: FIG. 3. Graphical proof of the invariance of the dCS [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗

**Figure 2.** Figure 2: FIG. 2. Graphical proof of the invariance of the dCS [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗

**Figure 4.** Figure 4: (a). By appropriately organizing terms around each site, we can directly obtain a TPS representation with the local tensors carrying three virtual indices: Ψ(s) = ∑ αβγ ∏ j T s2 j αjβjγ j T s2 j+1 αj+1βjγ j−1 , T s2 j αjβjγ j = ∑ α ′ j γ ′ j P s2 j α ′ j βjγ ′ j Ωα ′ j αjΩγ ′ j γ j , T s2 j+1 αj+1βjγ j−1 = ∑ β ′ j P s2 j+1 αj+1β ′ j γ j−1 Ω˜ βjβ ′ j . (3.15) To cast the wavefunction in the NQS representat… view at source ↗

**Figure 5.** Figure 5: FIG. 5. Graphical proof of the invariance of the dCS [PITH_FULL_IMAGE:figures/full_fig_p007_5.png] view at source ↗

**Figure 6.** Figure 6: FIG. 6. Graphical proof of the invariance of the dCS [PITH_FULL_IMAGE:figures/full_fig_p008_6.png] view at source ↗

**Figure 7.** Figure 7: FIG. 7 [PITH_FULL_IMAGE:figures/full_fig_p010_7.png] view at source ↗

read the original abstract

The $\mathbb{Z}_N$ cluster-state wavefunction, a paradigmatic example of symmetry-protected topological (SPT) order with $\mathbb{Z}_N \times \mathbb{Z}_N$ symmetry, is expressed in various equivalent ways. We identify the projector-based scheme called the $P$-representation as the efficient way to express cluster and dipolar cluster state's wavefunctions. Employing the restricted Boltzmann machine scheme to re-write the interaction matrix in the $P$-representation in terms of neural weight matrices allows us to develop the neural quantum state (NQS) and the matrix product state (MPS) representations of the same state. The NQS and MPS representations differ only in the way the weight matrices are split and grouped together in a matrix product. For both $\mathbb{Z}_N$ cluster and dipolar cluster states, we derive in closed form the weight function $W(s,h)$ that couples physical spins $s$ to hidden variables $h$, generalizing the previous construction for $Z_2$ cluster states to $\mathbb{Z}_N$. For the dipolar cluster state protected by two charge and two dipole symmetries, the procedure we have developed leads to the tensor product state (TPS) representation of the wavefunction where each local tensor carries three virtual indices connecting a given site to two nearest neighbors and one further neighbor. We benchmark the resulting TPS construction against conventional MPS representation using density-matrix renormalization group simulations and argue that the TPS could offer a more efficient representation for some modulated SPT states. As a by-product of the investigation, we generalize the previous $Z_2$ matrix product operator construction of the Kramers-Wannier (KW) operator to $\mathbb{Z}_N$ and interprets it as the dipolar generalization of the discrete Fourier transform on $\mathbb{Z}_N$ variables. The new interpretation naturally explains why the KW map is non-invertible.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

This paper supplies explicit closed-form NQS weights and a TPS representation for Z_N cluster and dipolar SPT states via direct algebraic extension of the projector form, without fitting.

read the letter

The main thing to know is that the authors derive closed-form weight functions W(s,h) for the Z_N cluster state and a three-index TPS for the dipolar version, both obtained by regrouping the known P-representation through the restricted Boltzmann machine mapping. They also extend the Kramers-Wannier operator to Z_N and read it as a dipolar discrete Fourier transform, which accounts for its non-invertibility. DMRG runs are used to compare the TPS against ordinary MPS and to argue for better scaling on modulated states. These steps are exact algebraic rearrangements rather than optimizations, so the topological content carries over directly from the starting projector. The TPS construction incorporates the required virtual indices for the two charge and two dipole symmetries in a natural way. This is useful technical work that fills in the arbitrary-N case and gives a concrete tensor-network ansatz where only the Z2 version existed before. The benchmarks add some evidence for the efficiency claim on modulated SPT states. The soft spots are modest. The efficiency edge of the TPS is demonstrated in specific DMRG comparisons but is not shown to be universal across all modulations or sizes, so the practical gain will depend on the target system. For large N the hidden-variable weights could grow cumbersome, though the paper does not test that regime. The derivations themselves look reproducible once the explicit formulas are written out. This is aimed at people who simulate SPT phases with tensor networks or neural states and need ready-made ansatze for Z_N generalizations. A reader working on cold-atom or quantum-magnet realizations of these phases would get concrete tools. It deserves a serious referee because the constructions are grounded in exact mappings, the benchmarks are present, and the generalizations are non-routine. I would send it for review.

Referee Report

0 major / 2 minor

Summary. The manuscript derives exact closed-form expressions for the weight function W(s,h) in the P-representation of Z_N cluster and dipolar-cluster SPT states, generalizing prior Z_2 constructions. These yield equivalent neural quantum state (NQS), matrix product state (MPS), and tensor product state (TPS) representations via algebraic regrouping of weight matrices, with the TPS for the dipolar case using three virtual indices to encode the two charge and two dipole symmetries. The work benchmarks the TPS against conventional MPS using DMRG simulations to argue for efficiency advantages in modulated SPT states and generalizes the Kramers-Wannier operator to Z_N, reinterpreting it as a dipolar discrete Fourier transform that explains its non-invertibility.

Significance. If the closed-form derivations hold as stated, the paper provides parameter-free, exact representations that unify projector, neural, and tensor-network descriptions of these SPT states. The explicit algebraic constructions and DMRG benchmarks constitute reproducible strengths that could aid numerical studies of modulated topological phases. The KW-operator generalization adds interpretive value by linking it to symmetry structure.

minor comments (2)

[Benchmarking and efficiency discussion] The DMRG benchmarking section should include explicit quantitative metrics (e.g., bond-dimension scaling or CPU-time comparisons) to support the efficiency claim for the TPS representation over MPS.
[TPS construction for dipolar cluster state] Clarify the precise range of N for which the closed-form W(s,h) remains valid without additional constraints, and confirm that the TPS virtual indices exactly capture the full symmetry group without truncation.

Simulated Author's Rebuttal

0 responses · 0 unresolved

We thank the referee for their careful reading of the manuscript and for the positive evaluation. We appreciate the recognition of the closed-form derivations for the Z_N cluster and dipolar-cluster SPT states, the unification of projector, neural, and tensor-network representations, the DMRG benchmarks, and the reinterpretation of the generalized Kramers-Wannier operator. Since the referee recommends minor revision but does not specify particular issues in the major comments section, we will review the manuscript for any typographical or minor clarifications in the revised version.

Circularity Check

0 steps flagged

No significant circularity: exact algebraic generalizations from projector form

full rationale

The paper begins with the known projector-based P-representation of the Z_N cluster and dipolar-cluster states. It then applies the standard restricted Boltzmann machine scheme to rewrite the interaction matrix, yielding closed-form weight functions W(s,h) via direct algebraic generalization of the Z_2 case. NQS, MPS, and TPS representations are obtained by exact regrouping and splitting of weight matrices into matrix products or tensors with the required virtual indices; no parameters are fitted to data or optimized variationally. DMRG benchmarks serve as independent numerical checks. The KW-operator extension is likewise a direct algebraic generalization reinterpreted as a dipolar discrete Fourier transform. No derivation step reduces to a self-definition, a fitted input renamed as prediction, or a load-bearing self-citation whose validity depends on the present work. The constructions remain self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

The central claims rest on standard assumptions of SPT physics (symmetry protection, projector definitions of cluster states) and tensor-network methods (MPS/TPS equivalence under certain splittings); no free parameters or invented entities are introduced beyond the known Z_N symmetry group.

axioms (2)

domain assumption The Z_N cluster state is exactly representable by a local projector onto the symmetric subspace defined by the Z_N x Z_N symmetry.
Invoked when identifying the P-representation as the starting point for all subsequent mappings.
domain assumption The restricted Boltzmann machine can exactly reproduce the interaction matrix of the cluster state when weights are chosen appropriately.
Underlying the derivation of closed-form W(s,h).

pith-pipeline@v0.9.0 · 5674 in / 1644 out tokens · 28208 ms · 2026-05-10T17:43:30.162537+00:00 · methodology

discussion (0)

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

IndisputableMonolith/Cost/FunctionalEquation.lean; Foundation/AbsoluteFloorClosure.lean; Foundation/AlexanderDuality.lean washburn_uniqueness_aczel; reality_from_one_distinction; alexander_duality_circle_linking unclear
We identify the projector-based scheme called the P-representation... derive in closed form the weight function W(s,h)... Ω=WW^t... TPS representation... KW operator as dipolar discrete Fourier transform
IndisputableMonolith/Cost.lean Jcost definition; costAlphaLog echoes
W(s,h)=κ^{-1/2} ω^{a h^2 + b s^2 + c s h} with a=N^{-1}/4, b=N^{-1}/2, c=-1 (App. B)

Reference graph

Works this paper leans on

44 extracted references · 11 canonical work pages · 1 internal anchor

[1]

kqv86ZV8yFHoK0jDgxMvsw3mgnA=

with the Hamiltonian H=− ∑ j Z2j−1 X2j Z−2 2j+1 Z2j+3 −∑ j Z2j−2 Z−2 2j X2j+1 Z2j+2 +h.c., (3.10) which is is symmetric under two sets of charge and dipole symmetries C1 = ∏ j X2j−1 ,C 2 = ∏ j X2j D1 = ∏ j (X2j−1 ) j,D 2 = ∏ j (X2j ) j.(3.11) To distinguish this from the previous dipolar CS, we refer to the new model as dCS II. The ground-state wavefuncti...

2024
[2]

Carleo and M

G. Carleo and M. Troyer, Solving the Quantum Many-Body Problem with Artificial Neural Networks, Science355, 602 (2017)

2017
[3]

L. L. Viteritti, R. Rende, and F. Becca, Transformer Variational Wave Functions for Frustrated Quantum Spin Systems, Phys. Rev. Lett.130, 236401 (2023)

2023
[4]

D.-L. Deng, X. Li, and S. Das Sarma, Machine learning topo- logical states, Phys. Rev. B96, 195145 (2017)

2017
[5]

D.-L. Deng, X. Li, and S. Das Sarma, Quantum Entanglement in Neural Network States, Phys. Rev. X7, 021021 (2017), arXiv:1701.04844 [cond-mat.dis-nn]

work page Pith review arXiv 2017
[6]

Gao and L.-M

X. Gao and L.-M. Duan, Efficient representation of quantum many-body states with deep neural networks, Nature Commu- nications8, 662 (2017)

2017
[7]

Glasser, N

I. Glasser, N. Pancotti, M. August, I. D. Rodr´ıguez, and I. Cirac, Neural-network quantum states, string-bond states, and chiral topological states, Phys. Rev. X8, 011006 (2018)

2018
[8]

S. R. Clark, Unifying neural-network quantum states and cor- relator product states via tensor networks, J. Phys. A: Math. Theor.51, 135301 (2018)

2018
[9]

J. Chen, S. Cheng, H. Xie, L. Wang, and T. Xiang, Equivalence of restricted Boltzmann machines and tensor network states, Phys. Rev. B97, 085104 (2018)

2018
[10]

Jia, Y .-H

Z.-A. Jia, Y .-H. Zhang, Y .-C. Wu, L. Kong, G.-C. Guo, and G.- P. Guo, Efficient machine-learning representations of a surface code with boundaries, defects, domain walls, and twists, Phys. Rev. A99, 012307 (2019)

2019
[11]

Z.-A. Jia, L. Wei, Y .-C. Wu, G.-C. Guo, and G.-P. Guo, Entan- glement area law for shallow and deep quantum neural network states, New Journal of Physics22, 053022 (2020)

2020
[12]

Sharir, A

O. Sharir, A. Shashua, and G. Carleo, Deep autoregressive mod- els for the efficient variational simulation of many-body quan- tum systems, Phys. Rev. Lett.124, 020503 (2020)

2020
[13]

Hibat-Allah, M

M. Hibat-Allah, M. Ganahl, R. Melko, and J. Carrasquilla, Re- current neural network wave functions, Phys. Rev. Research2, 023358 (2020)

2020
[14]

Carrasquilla, Machine learning for quantum matter, Ad- vances in Physics: X5, 1797528 (2020)

J. Carrasquilla, Machine learning for quantum matter, Ad- vances in Physics: X5, 1797528 (2020)

2020
[15]

Vivas and J

C. Vivas and J. Madro˜nero, Neural-Network Quantum States: A Systematic Review, arXiv preprint (2022), arXiv:2204.12966 [quant-ph]

work page arXiv 2022
[16]

Lange, A

H. Lange, A. Van de Walle, A. Abedinnia, and A. Bohrdt, From architectures to applications: a review of neural quantum states, Quantum Science and Technology9, 040501 (2024)

2024
[17]

P. Chen, B. Yan, and S. X. Cui, Representing arbitrary ground states of the toric code by a restricted Boltzmann machine, Phys. Rev. B111, 045101 (2025)

2025
[18]

Paul, Bound on entanglement in neural quantum states (2025), arXiv:2510.11797 [quant-ph]

N. Paul, Bound on entanglement in neural quantum states (2025), arXiv:2510.11797 [quant-ph]

work page arXiv 2025
[19]

Chen, Z.-C

X. Chen, Z.-C. Gu, and X.-G. Wen, Complete classification of one-dimensional gapped quantum phases in interacting spin systems, Phys. Rev. B84, 235128 (2011)

2011
[20]

Schuch, D

N. Schuch, D. P ´erez-Garc´ıa, and I. Cirac, Classifying quantum phases using matrix product states and projected entangled pair states, Phys. Rev. B84, 165139 (2011)

2011
[21]

Pollmann, A

F. Pollmann, A. M. Turner, E. Berg, and M. Oshikawa, Symme- try protection of topological phases in one-dimensional quan- tum spin systems, Phys. Rev. B85, 075125 (2012)

2012
[22]

J. H. Han, E. Lake, H. T. Lam, R. Verresen, and Y . You, Topo- logical quantum chains protected by dipolar and other modu- lated symmetries, Phys. Rev. B109, 125121 (2024)

2024
[23]

H. T. Lam, Classification of dipolar symmetry-protected topo- logical phases: Matrix product states, stabilizer Hamiltonians, and finite tensor gauge theories, Phys. Rev. B109, 115142 (2024), arXiv:2311.04962 [cond-mat.str-el]

work page arXiv 2024
[24]

J. Kim, Y . You, and J. H. Han, Noninvertible symmetry and topological holography for modulated SPT in one dimension, SciPost Phys.19, 110 (2025)

2025
[25]

Sun, J.-H

S. Sun, J.-H. Zhang, Z. Bi, and Y . You, Holographic View of Mixed-State Symmetry-Protected Topological Phases in Open Quantum Systems, PRX Quantum6, 020333 (2025)

2025
[26]

S. D. Pace, G. Delfino, H. T. Lam, and ¨Omer M. Aksoy, Gaug- ing modulated symmetries: Kramers-Wannier dualities and non-invertible reflections, SciPost Phys.18, 021 (2025)

2025
[27]

Saito, W

T. Saito, W. Cao, B. Han, and H. Ebisu, Matrix product state classification of one-dimensional multipole symmetry- protected topological phases, Physical Review B112, 10.1103/6txg-7hy7 (2025)

work page doi:10.1103/6txg-7hy7 2025
[28]

Bulmash, Defect Networks for Topological Phases Protected By Modulated Symmetries (2025), arXiv:2508.06604 [cond- mat.str-el]

D. Bulmash, Defect Networks for Topological Phases Protected By Modulated Symmetries (2025), arXiv:2508.06604 [cond- mat.str-el]

work page arXiv 2025
[29]

S. D. Pace, ¨Omer M. Aksoy, and H. T. Lam, Spacetime symmetry-enriched SymTFT: From LSM anomalies to modu- lated symmetries and beyond, SciPost Phys.20, 007 (2026). 18

2026
[30]

2026.arXiv e-prints:arXiv:2603.19189

A. Anakru, S. Srinivasan, L. Li, and Z. Bi, Matrix Product States for Modulated Symmetries: SPT, LSM, and Beyond (2026), arXiv:2603.19189 [cond-mat.str-el]

work page arXiv 2026
[31]

S.-Q. Ning, H. Ebisu, and H. T. Lam, Matrix Product States for Modulated Topological Phases: Crystalline Equiv- alence Principle and Lieb-Schultz-Mattis Constraints (2026), arXiv:2603.19381 [cond-mat.str-el]

work page internal anchor Pith review Pith/arXiv arXiv 2026
[32]

Raussendorf and H

R. Raussendorf and H. J. Briegel, A One-Way Quantum Com- puter, Phys. Rev. Lett.86, 5188 (2001)

2001
[33]

Maeda and T

J. Maeda and T. Oishi, N-ality symmetry and SPT phases in (1+1)d, Journal of High Energy Physics2025, 63 (2025)

2025
[34]

Yao, Lattice Translation Modulated Symmetries and TFTs (2025), arXiv:2510.03889 [cond-mat.str-el]

C.-Y . Yao, Lattice Translation Modulated Symmetries and TFTs (2025), arXiv:2510.03889 [cond-mat.str-el]

work page arXiv 2025
[35]

Seiberg, S

N. Seiberg, S. Seifnashri, and S.-H. Shao, Non-invertible sym- metries and LSM-type constraints on a tensor product Hilbert space, SciPost Phys.16, 154 (2024)

2024
[36]

Seifnashri and S.-H

S. Seifnashri and S.-H. Shao, Cluster State as a Noninvertible Symmetry-Protected Topological Phase, Phys. Rev. Lett.133, 116601 (2024)

2024
[37]

W. Cao, L. Li, and M. Yamazaki, Generating lattice non- invertible symmetries, SciPost Phys.17, 104 (2024)

2024
[38]

Aasen, R

D. Aasen, R. S. K. Mong, and P. Fendley, Topological defects on the lattice: I. The Ising model, Journal of Physics A: Mathe- matical and Theoretical49, 354001 (2016)

2016
[39]

Shao, What’s done cannot be undone: Tasi lectures on non-invertible symmetries (2024), arXiv:2308.00747 [hep-th]

S.-H. Shao, What’s done cannot be undone: TASI lectures on non-invertible symmetries (2024), arXiv:2308.00747 [hep-th]

work page arXiv 2024
[40]

Bhardwaj, S

L. Bhardwaj, S. Schafer-Nameki, and J. Yan, Generalized global symmetries and non-invertible symmetries, Communi- cations in Mathematical Physics393, 119 (2022)

2022
[41]

More precisely, the total dipole moment must be zero modulus N, assuming that the length of the chainLis divisible byN
[42]

J. Kim, M. Kim, N. Kawashima, J. H. Han, and H.-Y . Lee, Con- struction of variational matrix product states for the Heisenberg spin-1 chain, Phys. Rev. B102, 085117 (2020)

2020
[43]

Zaklama, M

T. Zaklama, M. Geier, and L. Fu, Large Electron Model: A Uni- versal Ground State Predictor (2026), arXiv:2603.02346 [cond- mat.str-el]

work page arXiv 2026
[44]

J. Kim, J. Y . Lee, and J. H. Han, Gauging modulated symme- tries via multiple gauge symmetry operators and adaptive quan- tum circuits, Phys. Rev. B113, 085139 (2026)

2026