arxiv: 2604.14067 · v1 · submitted 2026-04-15 · 🌀 gr-qc · hep-th· physics.comp-ph

Recognition: unknown

Finding and characterising physical states of Euclidean Abelianized loop quantum gravity using neural quantum states

Hanno Sahlmann , Waleed Sherif

Authors on Pith no claims yet

Pith reviewed 2026-05-10 12:41 UTC · model grok-4.3

classification 🌀 gr-qc hep-thphysics.comp-ph

keywords loop quantum gravityneural quantum statesvariational Monte CarloHamiltonian constraintphysical statesEuclidean gravityfinite graphAbelianized model

0 comments

The pith

Variational neural states identify distinct solution families for the Hamiltonian constraint and its adjoint in a truncated Euclidean loop quantum gravity model.

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

The paper applies variational Monte Carlo sampling with neural network representations of quantum states to search for approximate physical states on the complete graph with five vertices in an Abelianized version of four-dimensional Euclidean loop quantum gravity. It examines the Hamiltonian constraint in one specific ordering, the adjoint operator, and their sum, while imposing successive cutoffs on the available degrees of freedom. Distinct families of approximate solutions appear depending on the operator: one family remains flat on minimal loops, carries non-vanishing volume, and shows charges that spread out with larger cutoffs, while the adjoint family stays normalizable, develops non-trivial charge correlations, and lies in the kernel of the volume operator. The symmetrized operator yields mixed quasi-solutions. A reader would care because the long-standing problem of constructing explicit physical states in loop quantum gravity has few concrete examples, and this approach offers a numerical route to characterize them through correlation functions and geometric quantities while connecting finite truncations to the continuum.

Core claim

The variational optimisation selects distinct solution families for the Hamiltonian constraint in the proposed ordering and for its adjoint across several cutoffs on the kinematical degrees of freedom. The family selected for the constraint is flat on all minimal loops, exhibits non-vanishing volume expectation values, and shows edge charges that delocalise with increasing cutoff, pointing to approximations of non-normalisable solutions in the kinematical inner product. The family for the adjoint is normalisable in the inner product, displays non-trivial charge correlations, lies in the kernel of the volume operator, and lacks flatness on minimal loops. The sum of the two operators proves to

What carries the argument

Neural network ansatz for the quantum state wavefunction on the graph, variationally optimized by Monte Carlo sampling to drive the expectation value of the squared constraint operator toward zero.

If this is right

Edge-charge marginals in the constraint family spread further with each increase in cutoff, supporting their role as approximations to non-normalizable continuum states.
The adjoint family remains normalizable and shows persistent non-trivial charge correlations across the cutoffs examined.
Quasi-solutions obtained from the symmetrized operator combine flatness on minimal loops with partial volume suppression.
One- and two-point correlation functions together with area and volume observables cleanly separate the two families.
Solutions found on the finite graph furnish concrete indications of how truncated physical states relate to solutions in the full continuum theory.

Where Pith is reading between the lines

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

If the separation into two families survives on larger graphs, different orderings of the constraint may select distinct physical sectors rather than approximations to a single set of states.
The charge delocalization trend implies that any exact continuum solutions for the constraint family would have diverging norm, which could be checked by tracking the norm scaling in successive truncations.
Applying the same variational search to additional constraints or to non-Abelian models could test whether similar families appear when the Abelian approximation is relaxed.
The reported links between truncated and continuum solutions suggest that geometric observables computed on the graph may approach continuum values in a controlled way as the cutoff is removed.

Load-bearing premise

The neural network is expressive enough to represent the relevant physical states and the optimization procedure converges to representative minima without significant bias from architecture or random initialization.

What would settle it

An exact computation of the kernel on the graph with the smallest cutoff that produces no states matching either variational family, or a demonstration that volume expectation values for the first family vanish, would falsify the separation into distinct families.

read the original abstract

We study physical (near-kernel of constraints) states of 4-d Euclidean loop quantum gravity in Smolin's weak coupling limit on the complete graph $K_5$ using variational Monte Carlo with neural network quantum states. We investigate the Hamilton constraint $\hat{H}$ in the ordering proposed by Thiemann, as well as $\hat{H}^\dagger$ and $\hat{H}+\hat{H}^\dagger$. We find that the variational optimisation selects distinct solution families for $\hat{H}$ and $\hat{H}^\dagger$ across several considered cutoffs on the kinematical degrees of freedom. The solution family of $\hat{H}$ is flat on all minimal loops and has non-vanishing volume expectation values. Its edge-charge marginals delocalise with increasing cutoff, which indicates they are approximations to solutions that are non-normalisable in the kinematical inner product. The solution family for $\hat{H}^\dagger$ is normalisable, shows non-trivial charge correlations, lies in the kernel of volume and is not flat. $\hat{H}+\hat{H}^\dagger$ turns out to be much harder to solve and yields quasi-solutions combining features of both previous families. We characterise all solutions using chromaticity 1- and 2-point functions, minimal loop holonomies, geometric area and volume observables and show that the two families can be interpreted as, on the one hand, a family of states close to the Ashtekar-Lewandowski vacuum and the Dittrich-Geiller vacuum with some numerical noise on the other hand. We also present some results that link solutions of the truncated theory to solutions of the continuum theory.

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

Neural quantum states give a workable numerical handle on approximate physical states for the Hamiltonian constraint in a K5 truncation of Abelianized Euclidean LQG, but the optimization needs more checks to confirm the families are robust.

read the letter

The paper shows that neural quantum states plus variational Monte Carlo can locate approximate solutions to the Hamiltonian constraint on the K5 graph in Abelianized Euclidean LQG. They recover two separate families, one for Thiemann's ordering of Ĥ and one for Ĥ†, and map them to noisy versions of the Ashtekar-Lewandowski and Dittrich-Geiller vacua using chromaticity functions, holonomies, and volume observables. The Ĥ family is flat on minimal loops with non-zero volume; the Ĥ† family is normalizable with charge correlations and zero volume. The sum Ĥ + Ĥ† is harder and produces mixed states. This is the first time neural ansatzes have been tried on this constraint in the literature they cite, and the concrete characterization on a small graph is useful because analytic solutions remain out of reach for realistic truncations. The work is honest about the finite cutoff and presents the observables clearly. The main soft spot is numerical reliability. No error bars, multiple-seed statistics, or architecture ablations are described, so it is not yet clear whether the distinct families survive changes in initialization or network size or whether the optimizer consistently reaches representative minima. The finite K5 results are treated as indicative for the continuum without explicit scaling tests, which leaves that link tentative. The paper is for people working on numerical or variational methods inside loop quantum gravity. Anyone already using neural states or small-graph truncations will find the technique and the specific observables worth examining. I would send it for peer review. The idea is grounded and the execution shows real effort; referees can usefully press on the optimization diagnostics and the cutoff extrapolation.

Referee Report

2 major / 2 minor

Summary. The manuscript applies variational Monte Carlo with neural quantum states to approximate physical states (near-kernel of the constraints) in 4-d Euclidean Abelianized loop quantum gravity on the complete graph K5. It studies the Thiemann-ordered Hamiltonian constraint Ĥ, its adjoint Ĥ†, and their sum under several kinematical cutoffs. The optimization is reported to select distinct solution families: Ĥ yields states flat on minimal loops with non-vanishing volume expectations and delocalizing edge charges (interpreted as noisy approximations to the Ashtekar-Lewandowski vacuum); Ĥ† yields normalizable states with non-trivial charge correlations, zero volume, and non-flat holonomies (close to the Dittrich-Geiller vacuum); Ĥ + Ĥ† produces quasi-solutions mixing features of both. All families are characterized via chromaticity 1- and 2-point functions, minimal-loop holonomies, geometric area/volume observables, with some discussion linking truncated solutions to the continuum.

Significance. If the numerical results are robust, the work demonstrates a viable route for using neural-network ansatzes to explore the physical sector of loop quantum gravity on finite graphs, yielding concrete, characterizable approximate solutions that connect to analytically known vacua. The multi-observable characterization (holonomies, volumes, charge correlations) and the explicit comparison of orderings provide a useful template for future variational studies in quantum gravity. The approach also highlights practical challenges in solving the symmetrized constraint.

major comments (2)

[Section 4 (Numerical Results)] Section 4 (Numerical Results): The central claim that the variational optimization selects distinct, reproducible solution families for Ĥ versus Ĥ† rests on the reported expectation values and marginals. However, the manuscript supplies no error bars, no convergence diagnostics (e.g., loss curves or variance across epochs), and no statistics from multiple random seeds or initializations. Without these, it is impossible to determine whether the observed distinctions (flatness, volume non-vanishing, charge delocalization) reflect global or representative minima rather than architecture- or initialization-dependent local minima.
[Section 5 (Interpretation and Continuum Limit)] Section 5 (Interpretation and Continuum Limit): The finite-K5 cutoff results are interpreted as approximations to the Ashtekar-Lewandowski and Dittrich-Geiller vacua and as indicative of continuum behavior. Yet no explicit scaling analysis with cutoff removal is presented, nor are limit checks shown for key quantities (e.g., how volume expectations or charge delocalization behave as the cutoff is taken to infinity). This scaling step is load-bearing for the physical interpretation and the claimed link to continuum solutions.

minor comments (2)

The abstract and results sections refer to 'some numerical noise' when comparing to exact vacua; a quantitative measure (e.g., L2 distance to the known vacuum states or overlap diagnostics) would make the closeness claim more precise.
Notation for the neural-network parameters and the precise definition of the cutoffs on kinematical degrees of freedom could be collected in a single table or appendix for easier reproducibility.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the detailed and constructive report. The two major comments highlight important aspects of numerical robustness and physical interpretation that we will address in a revised manuscript. Below we respond point by point.

read point-by-point responses

Referee: [Section 4 (Numerical Results)] Section 4 (Numerical Results): The central claim that the variational optimization selects distinct, reproducible solution families for Ĥ versus Ĥ† rests on the reported expectation values and marginals. However, the manuscript supplies no error bars, no convergence diagnostics (e.g., loss curves or variance across epochs), and no statistics from multiple random seeds or initializations. Without these, it is impossible to determine whether the observed distinctions (flatness, volume non-vanishing, charge delocalization) reflect global or representative minima rather than architecture- or initialization-dependent local minima.

Authors: We agree that the absence of error bars, convergence diagnostics, and multi-seed statistics weakens the presentation of the numerical results. In the revised manuscript we will add (i) error bars estimated from the Monte Carlo sampling variance on all reported expectation values, (ii) representative loss curves and variance plots across training epochs, and (iii) summary statistics (means and standard deviations) from at least five independent runs with different random seeds and initial network weights. These additions will allow readers to assess the reproducibility of the two distinct solution families. revision: yes
Referee: [Section 5 (Interpretation and Continuum Limit)] Section 5 (Interpretation and Continuum Limit): The finite-K5 cutoff results are interpreted as approximations to the Ashtekar-Lewandowski and Dittrich-Geiller vacua and as indicative of continuum behavior. Yet no explicit scaling analysis with cutoff removal is presented, nor are limit checks shown for key quantities (e.g., how volume expectations or charge delocalization behave as the cutoff is taken to infinity). This scaling step is load-bearing for the physical interpretation and the claimed link to continuum solutions.

Authors: The manuscript already contains some explicit comparisons across the considered kinematical cutoffs (charge delocalization increasing with cutoff, persistence of flatness on minimal loops, and volume expectations remaining non-zero or zero). We acknowledge, however, that a more systematic scaling study with larger cutoffs would strengthen the continuum interpretation. In revision we will expand Section 5 with additional plots showing the cutoff dependence of the key observables (volume, charge correlations, minimal-loop holonomies) and will add a clearer discussion of the limitations imposed by the finite-graph truncation, while retaining the existing links to the Ashtekar-Lewandowski and Dittrich-Geiller vacua. revision: partial

Circularity Check

0 steps flagged

No circularity: results from direct numerical variational optimization

full rationale

The paper performs variational Monte Carlo optimization of a neural-network ansatz to minimize a constraint-violation functional on the finite K5 graph for the Thiemann-ordered Hamiltonian and its adjoint. The reported solution families (flat minimal-loop holonomies with non-zero volume for Ĥ; normalizable states with zero volume for Ĥ†) are direct outputs of this minimization rather than quantities defined in terms of the inputs or prior self-citations. No load-bearing step reduces by the paper's own equations to a fitted parameter renamed as a prediction, nor does any central claim rest on a uniqueness theorem or ansatz imported solely via self-citation. Post-hoc comparison to the Ashtekar-Lewandowski and Dittrich-Geiller vacua is interpretive and does not enter the derivation chain. The approach is therefore self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

2 free parameters · 2 axioms · 0 invented entities

The central claim rests on the validity of the Abelianized weak-coupling model, the Thiemann ordering of the constraint, and the assumption that the neural ansatz can faithfully represent the relevant states without introducing uncontrolled bias.

free parameters (2)

Neural network weights and biases
Variational parameters optimized to minimize constraint violation; their number grows with cutoff and network depth.
Kinematical cutoff
Several discrete cutoffs on degrees of freedom are chosen by hand and affect which family is selected.

axioms (2)

domain assumption Thiemann ordering of the Hamiltonian constraint operator
Invoked when defining Ĥ and Ĥ†.
domain assumption Smolin weak-coupling limit of Euclidean LQG
Defines the model in which the constraints are studied.

pith-pipeline@v0.9.0 · 5605 in / 1359 out tokens · 57438 ms · 2026-05-10T12:41:50.767505+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

45 extracted references · 41 canonical work pages

[1]

Loop quantum gravity,

C. Rovelli, “Loop quantum gravity,” Living Rev. Rel.1, 1 (1998) doi:10.12942/lrr-1998-1 [arXiv:gr- qc/9710008 [gr-qc]]

work page doi:10.12942/lrr-1998-1 1998
[2]

Introduction to Modern Canonical Quantum General Relativity,

T. Thiemann, “Introduction to Modern Canonical Quantum General Relativity,” [arXiv:gr- qc/0110034 [gr-qc]]

work page arXiv
[3]

Thiemann, Modern Canonical Quantum Gen- eral Relativity, Cambridge University Press, 2007, ISBN 978-0-511-75568-2, 978-0-521-84263-1 doi:10.1017/CBO9780511755682

T. Thiemann, “Modern Canonical Quantum General Relativity,” Cambridge University Press, 2007, ISBN 978-0-511-75568-2, 978-0-521-84263-1 doi:10.1017/CBO9780511755682

work page doi:10.1017/cbo9780511755682 2007
[4]

Ashtekar and J

A. Ashtekar and J. Lewandowski, “Background independent quantum gravity: A Status report,” Class. Quant. Grav.21(2004), R53 doi:10.1088/0264-9381/21/15/R01 [arXiv:gr-qc/0404018 [gr- qc]]

work page doi:10.1088/0264-9381/21/15/r01 2004
[5]

Ashtekar, New Variables for Classical and Quan- tum Gravity, Phys

A. Ashtekar, “New Variables for Classical and Quantum Gravity,” Phys. Rev. Lett.57(1986), 2244-2247 doi:10.1103/PhysRevLett.57.2244

work page doi:10.1103/physrevlett.57.2244 1986
[6]

Real Ashtekar variables for Lorentzian signature space times,

J. F. Barbero G., “Real Ashtekar variables for Lorentzian signature space times,” Phys. Rev. D51 (1995), 5507-5510 doi:10.1103/PhysRevD.51.5507 [arXiv:gr-qc/9410014 [gr-qc]]

work page doi:10.1103/physrevd.51.5507 1995
[7]

Spin network states in gauge theory,

J. C. Baez, “Spin network states in gauge theory,” Adv. Math.117(1996), 253-272 doi:10.1006/aima.1996.0012 [arXiv:gr-qc/9411007 [gr-qc]]

work page doi:10.1006/aima.1996.0012 1996
[8]

Quantum theory of geometry. 1: Area operators,

A. Ashtekar and J. Lewandowski, “Quantum theory of geometry. 1: Area operators,” Class. Quant. Grav.14(1997), A55-A82 doi:10.1088/0264-9381/14/1A/006 [arXiv:gr-qc/9602046 [gr-qc]]

work page doi:10.1088/0264-9381/14/1a/006 1997
[9]

Discreteness of area and volume in quantum gravity,

C. Rovelli and L. Smolin, “Discreteness of area and volume in quantum gravity,” Nucl. Phys. B442 (1995), 593-619 [erratum: Nucl. Phys. B456(1995), 753-754] doi:10.1016/0550-3213(95)00150-Q [arXiv:gr-qc/9411005 [gr-qc]]

work page doi:10.1016/0550-3213(95)00150-q 1995
[10]

Quantum theory of geometry. II: Volume operators,

A. Ashtekar and J. Lewandowski, “Quantum theory of geometry. II: Volume operators,” Adv. Theor. Math. Phys.1(1997), 388-429 doi:10.4310/ATMP.1997.v1.n2.a8 [arXiv:gr-qc/9711031 [gr- qc]]

work page doi:10.4310/atmp.1997.v1.n2.a8 1997
[11]

Quantum spin dynamics (QSD),

T. Thiemann, “Quantum spin dynamics (QSD),” Class. Quant. Grav.15(1998), 839-873 doi:10.1088/0264-9381/15/4/011 [arXiv:gr-qc/9606089 [gr-qc]]

work page doi:10.1088/0264-9381/15/4/011 1998
[12]

Quantum spin dynamics (qsd). 2.,

T. Thiemann, “Quantum spin dynamics (qsd). 2.,” Class. Quant. Grav.15, 875-905 (1998) doi:10.1088/0264-9381/15/4/012 [arXiv:gr-qc/9606090 [gr-qc]]

work page doi:10.1088/0264-9381/15/4/012 1998
[13]

QSD 3: Quantum constraint algebra and physical scalar product in quantum general relativity,

T. Thiemann, “QSD 3: Quantum constraint algebra and physical scalar product in quantum general relativity,” Class. Quant. Grav.15(1998), 1207-1247 doi:10.1088/0264-9381/15/5/010 [arXiv:gr-qc/9705017 [gr-qc]]

work page doi:10.1088/0264-9381/15/5/010 1998
[14]

QSD 4: (2+1) Euclidean quantum gravity as a model to test (3+1) Lorentzian quantum gravity,

T. Thiemann, “QSD 4: (2+1) Euclidean quantum gravity as a model to test (3+1) Lorentzian quantum gravity,” Class. Quant. Grav.15(1998), 1249-1280 doi:10.1088/0264-9381/15/5/011 [arXiv:gr-qc/9705018 [gr-qc]]

work page doi:10.1088/0264-9381/15/5/011 1998
[15]

’Sum over surfaces’ form of loop quantum gravity,

M. P. Reisenberger and C. Rovelli, “’Sum over surfaces’ form of loop quantum gravity,” Phys. Rev. D56, 3490-3508 (1997) doi:10.1103/PhysRevD.56.3490 [arXiv:gr-qc/9612035 [gr-qc]]

work page doi:10.1103/physrevd.56.3490 1997
[16]

Symmetric scalar constraint for loop quantum gravity,

J. Lewandowski and H. Sahlmann, “Symmetric scalar constraint for loop quantum gravity,” Phys. Rev. D91(2015) no.4, 044022 doi:10.1103/PhysRevD.91.044022 [arXiv:1410.5276 [gr-qc]]

work page doi:10.1103/physrevd.91.044022 2015
[17]

Anomaly free quantum dynamics for Euclidean LQG,

M. Varadarajan, “Anomaly free quantum dynamics for Euclidean LQG,” doi:10.48550/arXiv.2205.10779 [arXiv:2205.10779 [gr-qc]]

work page doi:10.48550/arxiv.2205.10779
[18]

Euclidean LQG Dynamics: An Electric Shift in Perspective,

M. Varadarajan, “Euclidean LQG Dynamics: An Electric Shift in Perspective,” Class. Quant. Grav.38(2021) no.13, 135020 doi:10.1088/1361-6382/abfc2d [arXiv:2101.03115 [gr-qc]]

work page doi:10.1088/1361-6382/abfc2d 2021
[19]

A short review of loop quantum gravity,

A. Ashtekar and E. Bianchi, “A short review of loop quantum gravity,” Rep. Prog. Phys.84(2021) no.4, 042001 doi:10.1088/1361-6633/abed91 [arXiv:2104.04394 [gr-qc]]

work page doi:10.1088/1361-6633/abed91 2021
[20]

Taming Thiemann’s 54 Hamiltonian constraint in canonical loop quantum gravity: Reversibility, eigenstates, and graph- change analysis,

T. L. M. Guedes, G. A. Mena Marug´ an, M. M¨ uller and F. Vidotto, “Taming Thiemann’s 54 Hamiltonian constraint in canonical loop quantum gravity: Reversibility, eigenstates, and graph- change analysis,” Phys. Rev. D112(2025) no.2, 026024 doi:10.1103/hjdk-kdhk [arXiv:2412.20272 [gr-qc]]

work page doi:10.1103/hjdk-kdhk 2025
[21]

Computing the Graph-Changing Dynamics of Loop Quantum Gravity,

T. L. M. Guedes, G. A. Mena Marug´ an, F. Vidotto and M. M¨ uller, “Computing the Graph-Changing Dynamics of Loop Quantum Gravity,” Universe11(2025) no.12, 387 doi:10.3390/universe11120387 [arXiv:2412.20257 [gr-qc]]

work page doi:10.3390/universe11120387 2025
[22]

Time evolution in deparametrized models of loop quantum gravity,

M. Assanioussi, J. Lewandowski and I. M¨ akinen, “Time evolution in deparametrized models of loop quantum gravity,” Phys. Rev. D96(2017) no.2, 024043 doi:10.1103/PhysRevD.96.024043 [arXiv:1702.01688 [gr-qc]]

work page doi:10.1103/physrevd.96.024043 2017
[23]

Bouncing Universe in loop quantum gravity: full theory calculation,

M. Kisielowski, “Bouncing Universe in loop quantum gravity: full theory calculation,” Class. Quant. Grav.40(2023) no.19, 195025 doi:10.1088/1361-6382/acf271 [arXiv:2211.04440 [gr-qc]]

work page doi:10.1088/1361-6382/acf271 2023
[24]

Time evolution of semiclassical states in the one-vertex model of quantum-reduced loop gravity,

I. M¨ akinen, “Time evolution of semiclassical states in the one-vertex model of quantum-reduced loop gravity,” [arXiv:2604.00999 [gr-qc]]

work page arXiv
[25]

Towards quantum gravity with neural networks: solving the quantum Hamilton constraint of U(1) BF theory,

H. Sahlmann and W. Sherif, “Towards quantum gravity with neural networks: solving the quantum Hamilton constraint of U(1) BF theory,” Class. Quant. Grav.41(2024) no.22, 225014 doi:10.1088/1361-6382/ad84af [arXiv:2402.10622 [gr-qc]]

work page doi:10.1088/1361-6382/ad84af 2024
[26]

Carleo and M

G. Carleo and M. Troyer, “Solving the quantum many-body problem with artificial neural networks,” Science355(2017) no.6325, 602-606 doi:10.1126/science.aag2302 [arXiv:1606.02318 [cond-mat.dis-nn]]

work page doi:10.1126/science.aag2302 2017
[27]

Towards quantum gravity with neural networks: solving quantum Hamilton constraints of 3d Euclidean gravity in the weak coupling limit,

H. Sahlmann and W. Sherif, “Towards quantum gravity with neural networks: solving quantum Hamilton constraints of 3d Euclidean gravity in the weak coupling limit,” Class. Quant. Grav.41 (2024) no.21, 215006 doi:10.1088/1361-6382/ad7c14 [arXiv:2405.00661 [gr-qc]]

work page doi:10.1088/1361-6382/ad7c14 2024
[28]

The G(Newton) —>0 limit of Euclidean quantum gravity,

L. Smolin, “The G(Newton) —>0 limit of Euclidean quantum gravity,” Class. Quant. Grav.9 (1992), 883-894 doi:10.1088/0264-9381/9/4/007 [arXiv:hep-th/9202076 [hep-th]]

work page doi:10.1088/0264-9381/9/4/007 1992
[29]

Projective techniques and functional integration for gauge theories,

A. Ashtekar and J. Lewandowski, “Projective techniques and functional integration for gauge theories,” J. Math. Phys.36(1995), 2170-2191 doi:10.1063/1.531037 [arXiv:gr-qc/9411046 [gr- qc]]

work page doi:10.1063/1.531037 1995
[30]

Differential geometry on the space of connections via graphs and projective limits,

A. Ashtekar and J. Lewandowski, “Differential geometry on the space of connections via graphs and projective limits,” J. Geom. Phys.17(1995), 191-230 doi:10.1016/0393-0440(95)00028-G [arXiv:hep-th/9412073 [hep-th]]

work page doi:10.1016/0393-0440(95)00028-g 1995
[31]

The Phoenix Project: Master Constraint Programme for Loop Quantum Gravity,

T. Thiemann, “The Phoenix Project: Master Constraint Programme for Loop Quantum Gravity,” Class. Quant. Grav.23(2006), 2211-2247 doi:10.1088/0264-9381/23/7/002 [arXiv:gr-qc/0305080 [gr-qc]]

work page doi:10.1088/0264-9381/23/7/002 2006
[32]

Andersson and G

B. Dittrich and T. Thiemann, “Testing the Master Constraint Programme for Loop Quantum Gravity: I. General Framework,” Class. Quant. Grav.23(2006), 1025-1065 doi:10.1088/0264- 9381/23/4/001 [arXiv:gr-qc/0411138 [gr-qc]]

work page doi:10.1088/0264- 2006
[33]

Quantization of diffeomorphism invariant theories of connections with local degrees of freedom,

A. Ashtekar, J. Lewandowski, D. Marolf, J. M. Mour˜ ao and T. Thiemann, “Quantization of diffeomorphism invariant theories of connections with local degrees of freedom,” J. Math. Phys. 36(1995), 6456-6493 doi:10.1063/1.531252 [arXiv:gr-qc/9504018 [gr-qc]]

work page doi:10.1063/1.531252 1995
[34]

Covariant origin of the U(1) 3 model for Euclidean quantum gravity,

S. Bakhoda and T. Thiemann, “Covariant origin of the U(1) 3 model for Euclidean quantum gravity,” Class. Quant. Grav.39(2022) no.2, 025006 doi:10.1088/1361-6382/ac37a4 [arXiv:2011.00031 [gr-qc]]

work page doi:10.1088/1361-6382/ac37a4 2022
[35]

TheU(1) 3 Model of Euclidean Quantum Gravity

S. Bakhoda, “TheU(1) 3 Model of Euclidean Quantum Gravity”, Doc- toral thesis, Friedrich-Alexander-Universit¨ at Erlangen–N¨ urnberg (FAU), 2022, https://open.fau.de/handle/openfau/19853

2022
[36]

Effective dynamics of weak coupling loop quantum gravity,

G. Long and Y. Ma, “Effective dynamics of weak coupling loop quantum gravity,” Phys. Rev. D 105(2022), 044043 doi:10.1103/PhysRevD.105.044043 [arXiv:2111.11844 [gr-qc]]

work page doi:10.1103/physrevd.105.044043 2022
[37]

Geometrical Quantum Time in the U(1) 3 Model of Euclidean Quantum Gravity,

S. Bakhoda and Y. Ma, “Geometrical Quantum Time in the U(1) 3 Model of Euclidean Quantum Gravity,” Commun. Theor. Phys.77(2025), 055401 doi:10.1088/1572-9494/ad972b [arXiv:2411.19435 [gr-qc]]

work page doi:10.1088/1572-9494/ad972b 2025
[38]

Simultaneous approximation of multiple degenerate states using a single neural network 55 quantum state,

W. Sherif, “Simultaneous approximation of multiple degenerate states using a single neural network 55 quantum state,” Mach. Learn. Sci. Tech.7(2026) no.1, 015029 doi:10.1088/2632-2153/ae3e39 [arXiv:2509.02658 [quant-ph]]

work page doi:10.1088/2632-2153/ae3e39 2026
[39]

A new vacuum for Loop Quantum Gravity,

B. Dittrich and M. Geiller, “A new vacuum for Loop Quantum Gravity,” Class. Quant. Grav.32 (2015) no.11, 112001 doi:10.1088/0264-9381/32/11/112001 [arXiv:1401.6441 [gr-qc]]

work page doi:10.1088/0264-9381/32/11/112001 2015
[40]

neuraLQX: a high-performance simulations toolkit for loop quantum gravity,

W. Sherif, “neuraLQX: a high-performance simulations toolkit for loop quantum gravity,” version 1.1.1 (2026) [software, available at: http://github.com/waleed-sh/neuraLQX]

2026
[41]

SoftwareX10, 100311 (2019) https://doi.org/10.1016/j.softx.2019.100311

G. Carleo, K. Choo, D. Hofmann, J. E. T. Smith, T. Westerhout, F. Alet, E. J. Davis, S. Efthymiou, I. Glasser, S.-H. Lin, M. Mauri, G. Mazzola, C. B. Pereira and F. Vicentini, “NetKet: A machine learning toolkit for many-body quantum systems,” SoftwareX10(2019), 100311 doi:10.1016/j.softx.2019.100311

work page doi:10.1016/j.softx.2019.100311 2019
[42]

Vicentini, D

F. Vicentini, D. Hofmann, A. Szab´ o, D. Wu, C. Roth, C. Giuliani, G. Pescia, J. Nys, V. Vargas- Calder´ on, N. Astrakhantsev and G. Carleo, “NetKet 3: Machine Learning Toolbox for Many-Body Quantum Systems,” SciPost Phys. Codebases (2022), 7 doi:10.21468/SciPostPhysCodeb.7

work page doi:10.21468/scipostphyscodeb.7 2022
[43]

JAX: Composable Transformations of Python + NumPy Programs,

J. Bradbury, R. Frostig, P. Hawkins, M. J. Johnson, C. Leary, D. Maclaurin, G. Necula, A. Paszke, J. VanderPlas, S. Wanderman-Milne and Q. Zhang, “JAX: Composable Transformations of Python + NumPy Programs,” (2018)

2018
[44]

Flax: A Neural Network Library and Ecosystem for JAX,

J. Heek, A. Levskaya, A. Oliver, M. Ritter, B. Rondepierre, A. Steiner and M. van Zee, “Flax: A Neural Network Library and Ecosystem for JAX,” (2024)

2024
[45]

mpi4jax: Zero-copy MPI communication of JAX arrays,

D. H¨ afner and F. Vicentini, “mpi4jax: Zero-copy MPI communication of JAX arrays,” J. Open Source Softw.6(2021) no.65, 3419 doi:10.21105/joss.03419. 56 Appendices A. State visualisation in large Hilbert spaces Typical Hilbert space dimensions, even in the gauge invariant subspace, considered in this work make it prohibitive to conduct any visualisation o...

work page doi:10.21105/joss.03419 2021