Pith. sign in

REVIEW 3 major objections 4 minor 78 references

Deep learning spinfoam vertex amplitudes: the Euclidean Barrett-Crane model

T0 review · 3 major / 4 minor · reviewed 2026-08-15 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read A neural network can learn the spinfoam vertex amplitude from data.

desk verdict A genuine first demonstration that a tiny neural net can learn the zero/nonzero structure of the 10j symbol and generalize beyond its training cutoff; the regression half is weaker than its headline metrics suggest. read the letter →

arxiv 2505.03255 v2 pith:NNUASFBR submitted 2025-05-06 gr-qc hep-thphysics.comp-ph

Hanno Sahlmann , Waleed Sherif This is my paper · ORCID
classification gr-qchep-thphysics.comp-ph
keywords loopquantumgravityspinfoammodelsBarrett-Cranemodel10jsymbolvertexamplitudedeeplearningneuralnetworksurrogatesupervised
topics Quantum Gravity
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

This paper tries to establish that the central building block of a spinfoam quantum-gravity calculation, the vertex amplitude, can be learned from data with deep neural networks. Working with the Euclidean Barrett-Crane model, the authors split the task into deciding whether the amplitude vanishes and predicting its numerical size. A tiny classifier, fed only the ten spins, distinguishes zero from non-zero configurations with an F1 score of 0.9947 on spins beyond its training range, and a larger regressor predicts the squared amplitude within its training cutoff to a mean absolute percentage error of 2.76%. If the result holds, exact but expensive amplitude evaluations could be replaced or pre-filtered by fast neural surrogates in the low-spin regime, a step toward the computationally hard sums over histories in spinfoam gravity.

What carries the argument

The central object is the Riemannian 10j symbol $\{10j\}$, the SU(2)$\times$SU(2) spin-network evaluation that carries the Barrett-Crane vertex amplitude $A_v(j_f)=(\prod_f(2j_f+1)^k)\{10j\}$. The machinery that makes learning work is a two-stage supervised pipeline: a binary classifier, a one-hidden-layer MLP with 1537 parameters and ReLU activation fed the raw ten spin values, learns the zero versus non-zero pattern, and a regressor, a deeper MLP with GELU activation fed one-hot encodings of the ten spins, learns $\log((\{10j\})^2+\epsilon)$ to tame the wide dynamic range of the squared symbol. The one-hot encoding removes artificial ordinality among spin labels, the log target compresses the dynamic range, and cutoff-wise transfer learning lets the classifier be retrained with progressively smaller datasets while improving out-of-domain metrics.

What would settle it

Re-train the $j_{\max}=1.0$ regressor on the same 85% of non-zero configurations, hold out the remaining 15%, and compute MAPE and $R^2$ on that held-out set alone; if the held-out MAPE is far above 2.76% or $R^2$ drops substantially below 0.9986, the reported accuracy is interpolation over training data rather than prediction of unseen configurations.

Watch

Extended reading notes

Core claim

On the paper's own terms, the discovery is that the vertex amplitude of the Euclidean Barrett-Crane model, $A_v(j_f)=(\prod_f(2j_f+1)^k)\{10j\}$, is learnable: a 1537-parameter multilayer perceptron can classify configurations by whether $\{10j\}$ vanishes, maintaining an F1 score of 0.9947 when trained only up to $j_{\max}=2.0$ and tested at $j_{\max}=2.5$, while a larger regressor trained at $j_{\max}=1.0$ predicts $\log((\{10j\})^2+\epsilon)$ with a mean absolute percentage error of 2.76% and $R^2=0.9986$ within its training domain. The authors report that the classifier generalizes far outside its training cutoff and shows no catastrophic forgetting under transfer learning, whereas the regressor does not generalize beyond its training domain. Together, a classifier, a regressor, and the known sign and dimension factors form an Expert network that outputs the full amplitude.

Load-bearing premise

The load-bearing premise is that the regressor's reported errors are errors on configurations the network has not seen: only 85% of the non-zero configurations at a cutoff were used for training, while the reported metrics cover all non-zero configurations, so if those metrics include training points the proof-of-principle rests on interpolation rather than prediction.

Editorial extensions

If this is right

  • A fast, accurate classifier of zero versus non-zero 10j configurations could act as a pre-filter in spinfoam sums, discarding vanishing amplitudes before any expensive exact evaluation.
  • Within the trained low-spin cutoff, the regressor can serve as a surrogate for the exact 10j symbol, with typical errors near 2.8% at $j_{\max}=1.0$, making repeated amplitude evaluations much cheaper.
  • Because the Barrett-Crane vertex appears as a limit of more physical EPRL-like models, the same classification-plus-regression architecture is a template for surrogates of those amplitudes, with the Immirzi parameter added as an input.
  • The Expert combination of classifier, regressor, sign factor, and dimension factor yields the complete vertex amplitude $A_v(j_f)$ from raw spins, the ingredient needed to accelerate the full partition-function sum.
  • Small training sets suffice for the classification part: at $j_{\max}=2.0$ only a small fraction of all configurations were used, yet metrics on cutoffs beyond the training range remained high, isolating the data-hungry part of the problem to regression.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • If the classifier's out-of-domain generalization is genuine, the zero versus non-zero structure of the 10j symbol is likely controlled by simple combinatorial inequalities that the network has learned implicitly; extracting those rules symbolically could replace the classifier with a closed-form test.
  • The regressor's failure to extrapolate across cutoffs suggests that a practical surrogate should be built as a mixture of per-cutoff experts rather than a single global network, a testable next step the paper names as future work.
  • Because the 10j symbol is a graph evaluation, graph neural networks may learn the amplitude from fewer parameters than the MLP regressors used here; comparing parameter counts and accuracy on the same data would settle whether the architecture, not the target, limits efficiency.
  • The same two-stage setup could be applied to Lorentzian and EPRL spinfoam amplitudes once enough exact data can be generated, turning the proof-of-principle into a practical accelerator for physically relevant models.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 4 minor

Summary. The manuscript proposes a supervised-learning surrogate for the Euclidean Barrett-Crane vertex amplitude, focusing on the Riemannian 10j symbol. Exact training data are generated by enumerating all spin configurations for cutoffs jmax = 0.5, 1.0, 1.5, 2.0, 2.5, using a Python/numba port of the Christensen-Egan algorithm. The learning problem is split into a binary classification task (zero versus nonzero 10j) and a regression task for log((10j)^2). The classifier is a 1537-parameter MLP trained with transfer learning up to jmax = 2.0 and tested on all cutoffs up to 2.5, with the best out-of-cutoff result F1 = 0.9947 at jmax = 2.5. The regressor is a deeper MLP with one-hot encoded inputs, trained on 85% of the non-zero configurations at jmax = 1.0 and 1.5; Table 3 reports MAPE = 2.7587% and R2 = 0.9986 at jmax = 1.0, and MAPE = 4.1735% at jmax = 1.5. A meta 'Expert' network combining the classifier and regressor is described but not evaluated. Section 4 candidly lists limitations, including data scarcity, regressor overparameterization, and the regressor's failure to generalize beyond the training cutoff.

Significance. If the regression metrics are confirmed on held-out data, the paper provides a useful proof-of-principle: it would be the first demonstration, to my knowledge, that a neural surrogate can reproduce the Barrett-Crane vertex amplitude in the low-spin regime. The classifier result is the strongest part of the paper: the out-of-cutoff generalization at F1 = 0.9947 is difficult to explain by memorization and suggests practical value for pre-selecting nonzero configurations in spinfoam sums. The paper also has genuine strengths in methodology: fully enumerated exact training data, a clean transfer-learning protocol for the classifier, confusion matrices, and an unusually candid limitations section. The main caveat is that the regression evidence is currently in-sample: because the regressor has more parameters than training configurations at jmax = 1.0, the reported accuracy may measure interpolation on memorized data rather than prediction on unseen inputs. The paper's own Section 4 acknowledges that Params(R) exceeds the number of available states and that the regressor does not generalize across cutoffs, which makes the missing held-out evaluation load-bearing.

major comments (3)
  1. [Section 2.3.2 and Table 3] The regression metrics in Table 3 are reported on 'all non-zero configurations' after the regressor was trained on 85% of the same configurations, with no held-out split or per-split metrics described. At jmax = 1.0 the regressor has 350,093 parameters while the training set contains roughly 16,162 non-zero configurations, so the network is capable of memorizing the training data. The reported MAPE = 2.7587% and R2 = 0.9986 therefore do not yet establish predictive accuracy on unseen configurations. Since the proof-of-principle depends on the networks predicting configurations they were not trained on, please restate all regression metrics on the held-out 15% fraction (ideally with k-fold or repeated-seed results), and separate in-domain from out-of-domain test results.
  2. [Section 3, Table 2] The claim that the classifier demonstrates 'robust generalisation far outside the training domain' is supported by aggregate metrics over all configurations at the test cutoff jmax = 2.5, which include configurations whose spins all lie in the training range j <= 2.0. To make the extrapolation quantitative, report metrics restricted to configurations containing at least one spin in the interval (2.0, 2.5], or otherwise show that the aggregate F1 = 0.9947 is not dominated by configurations already representable in the training distribution.
  3. [Section 3.0.1] The meta 'Expert' network P(S) is described as producing the correct full vertex amplitude by combining the classifier, the regressor, the sign reconstruction from equation (6), and the dimension factor from equation (3), but it is never evaluated end-to-end. Since the abstract claims that 'the vertex amplitude is an object that can be learned from data,' the final product should be assessed on held-out configurations; otherwise the paper establishes only that the component classification and regression tasks are individually learnable, not that the composed amplitude prediction works.
minor comments (4)
  1. [Section 3, Table 1] The text states that the training dataset size at jmax = 2.0 is 1.17% of the available states, but the table gives Ntrain/Ns = 170356/9765625, which is approximately 1.74%; please correct the inconsistency.
  2. [Section 2.1] The notation sigma(S(jmax)) = sgn({10j}(S(jmax)))^2 is confusing: since ({10j})^2 is nonnegative, this quantity is a zero/nonzero indicator, not a sign. Please rename the label accordingly and avoid the phrase 'sign factor' for this binary target.
  3. [Section 2.2] The one-hot encoding justification says that spin values are categorical symbols rather than points on a meaningful real line, yet the classifier takes raw spin values as input. Please explain why raw numerical inputs are acceptable for classification while one-hot encoding is used for regression, or use a consistent encoding.
  4. [Section 2.3.2] In the definition of the Huber loss, the error e is written as the difference between two identical-looking expressions, with the hat on the predicted value missing in the second term. Please fix the notation so that e = y_log - yhat_log.

Circularity Check

1 steps flagged · score 6.0 of 10

Regression 'prediction' metrics are evaluated on a set that includes 85% of the training data, so the reported in-domain accuracy is partly a restatement of the fit rather than an out-of-sample test.

  1. fitted input called prediction [Section 2.3.2 (Regression metrics and protocol); Table 3; Figure 5; Section 4 (Discussion)]
    "For all cutoffs, the training data consisted of 85% of all available non-zero configurations S(jmax) at the current cutoff. ... [Table 3:] 'Evaluation metrics on all non-zero configurations for the regressors trained at different cutoffs.' ... 'with enough learnable parameters, one can fit any data. In this case, it is much faster to simply create a table of all possible amplitude values for all spins in the current cutoff.'"

    The regressor's headline metrics (MAPE 2.7587%, R2 0.9986 at jmax=1.0) are computed on 'all non-zero configurations', which includes the 85% of configurations used for training. No held-out-only regression metrics are reported, so the in-domain 'predictive accuracy' conflates memorization with generalization. With 350,093 parameters versus 16,162 training points, near-perfect in-sample performance is expected; the paper itself concedes that with enough parameters one can fit any data and that the regressors did not generalize beyond the training cutoff. Thus the claim that the regressor 'demonstrates high predictive accuracy in the domain it is trained on' is in large part a restatement of the fit, not evidence of prediction on unseen configurations.

full rationale

The paper's physics content is largely self-contained: the target amplitude Av(jf) = (product_f (2jf+1)^k) {10j} is taken from the literature and evaluated by an exact algorithm, and the networks are fit to those exact values, so there is no claim of deriving the amplitude from first principles. The authors' prior self-citations (refs. [15,16]) are background on neural networks for Hamilton constraints and are not load-bearing for the BC vertex learning. The classifier result is a genuine extrapolation: trained up to jmax=2.0 with small training fractions and tested up to jmax=2.5, it reaches F1=0.9947, so the central proof-of-principle has independent content. The significant circularity is the regression evaluation, where Table 3 and Figure 5 report metrics on 'all non-zero configurations' while 85% of those configurations were used for training, and no separate held-out split is described. Section 4 explicitly acknowledges the regressor is inefficient, has more parameters than available states, and does not generalize to higher cutoffs, confirming that the high in-domain metrics are partly a fit. This is a partial circularity in one of the two central 'prediction' claims, but not a full reduction of the paper's overall claim because the classifier provides a genuine out-of-sample result and the regression limitations are openly disclosed.

Assumptions & free parameters 4 free parameters · 4 assumptions · 0 invented entities

The central claim rests on exact training labels from the Christensen-Egan algorithm, the correctness of the BC vertex amplitude formula and sign relation, and the representativeness of small-cutoff enumeration. The neural network weights are free parameters fitted to the data; no new physical entities are introduced.

free parameters (4)
  • Classifier network weights = 1537 parameters, trained on classification data up to jmax=2.0
    The classifier function C(S) is defined entirely by these fitted weights; all reported classification accuracy metrics characterize this fitted function.
  • Regressor network weights = 350,093 parameters at jmax=1.0; 1,091,725 at jmax=1.5
    The regressor R(S) is defined by these fitted weights; its in-domain MAPE and R2 describe the fit, not an independent amplitude derivation.
  • Hyperparameters (learning rates, batch sizes, depth/width, weight decay, epochs) = Peak LR 1e-3 regressor, 1e-4 classifier; batch 256 regressor, 8 classifier; weight decay 1e-6; 200 epochs
    Selected via Optuna and manual choices; the reported performance is conditional on these choices and no sensitivity analysis is given.
  • Epsilon offset in regression log-target = 1e-26
    Small additive constant in log(({10j})^2 + epsilon) to handle zero values; hand-chosen and negligible for nonzero amplitudes, but part of the target definition.
assumptions (4)
  • standard math The Christensen-Egan algorithm [63] provides exact values of the Riemannian 10j symbols used as training labels.
    All classifier and regressor labels are generated from this algorithm; any error in it becomes an error in the learned amplitude.
  • domain assumption The BC vertex amplitude is Av = (product_f (2j_f+1)^k) {10j}, and the sign of the 10j symbol is recovered from equation (6).
    The Expert network reconstructs the physical amplitude using these formulas from references [61,62]; the paper does not rederive them.
  • domain assumption Full enumeration of spin configurations at small cutoffs produces a dataset representative enough for supervised training.
    The authors train on 75% down to 1% of configurations and assume the sampled subset supports learning; no theoretical guarantee is given.
  • domain assumption The metrics in Table 3, computed on all non-zero configurations, are treated as predictive accuracy rather than pure in-sample fit.
    This assumption is questionable because no held-out split is described; it is the weakest load-bearing premise in the regression evaluation.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Deep learning spinfoam vertex amplitudes: the Euclidean Barrett-Crane model." pith.science (2026). https://pith.science/paper/NNUASFBR

@misc{pith2026250503255,
  author       = {Pith},
  title        = {Pith review of: Deep learning spinfoam vertex amplitudes: the Euclidean Barrett-Crane model},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/NNUASFBR}},
  note         = {Machine review of arXiv:2505.03255}
}
read the original abstract

Spinfoam theories propose a well-defined path-integral formulation for quantum gravity and are hoped to provide the dynamics of loop quantum gravity. However, it is computationally hard to calculate spinfoam amplitudes. The well-studied Euclidean Barrett-Crane model provides an excellent setting for testing analytical and numerical tools to probe spinfoam models. We explore a data-driven approach to accelerating spinfoam computations by showing that the vertex amplitude is an object that can be learned from data using deep learning. We divide the learning process into a classification and a regression task: Two networks are independently engineered to decide whether the amplitude is zero or not and to predict the precise numerical value, respectively. The trained networks are tested with several accuracy measures. The classifier in particular demonstrates robust generalisation far outside the training domain, while the regressor demonstrates high predictive accuracy in the domain it is trained on. We discuss limitations, possible improvements, and implications for future work.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

78 extracted references · 18 canonical work pages

  1. [1]

    Loop quantum gravity,

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

    arXiv 1998

  2. [2]

    Modern canonical quantum general relativity,

    T. Thiemann, “Modern canonical quantum general relativity,” [arXiv:gr-qc/0110034 [gr-qc]]

    arXiv

  3. [3]

    Background independent quantum gravity: A Status report,

    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]]

    arXiv 2004

  4. [4]

    Fundamental structure of loop quantum gravity,

    M. Han, W. Huang and Y. Ma, “Fundamental structure of loop quantum gravity,” Int. J. Mod. Phys. D 16 (2007), 1397-1474 doi:10.1142/S0218271807010894 [arXiv:gr-qc/0509064 [gr-qc]]

    arXiv 2007

  5. [5]

    The Phoenix project: Master constraint program for loop quantum gravity,

    T. Thiemann, “The Phoenix project: Master constraint program for loop quantum gravity,” Class. Quant. Grav. 23 (2006), 2211-2248 doi:10.1088/0264-9381/23/7/002 [arXiv:gr-qc/0305080 [gr-qc]]

    arXiv 2006

  6. [6]

    Testing the master constraint programme for loop quantum gravity. I. General framework,

    B. Dittrich and T. Thiemann, “Testing the master constraint programme for loop quantum gravity. I. General framework,” Class. Quant. Grav. 23 (2006), 1025-1066 doi:10.1088/0264-9381/23/4/001 [arXiv:gr-qc/0411138 [gr-qc]]

    arXiv 2006

  7. [7]

    Quantum spin dynamics. VIII. The Master constraint,

    T. Thiemann, “Quantum spin dynamics. VIII. The Master constraint,” Class. Quant. Grav. 23 (2006), 2249-2266 doi:10.1088/0264-9381/23/7/003 [arXiv:gr-qc/0510011 [gr-qc]]

    arXiv 2006

  8. [8]

    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]]

    arXiv 1998

Show all 78 references
  1. [9]

    Quantum spin dynamics (qsd). 2.,

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

    1998 arXiv

  2. [10]

    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]]

    doi arXiv

  3. [11]

    Absence of singularity in loop quantum cosmology,

    M. Bojowald, “Absence of singularity in loop quantum cosmology,” Phys. Rev. Lett. 86 (2001), 5227-5230 doi:10.1103/PhysRevLett.86.5227 [arXiv:gr-qc/0102069 [gr-qc]]

    2001 arXiv

  4. [12]

    Quantum Nature of the Big Bang: Improved dynamics,

    A. Ashtekar, T. Pawlowski and P. Singh, “Quantum Nature of the Big Bang: Improved dynamics,” Phys. Rev. D 74 (2006), 084003 doi:10.1103/PhysRevD.74.084003 [arXiv:gr-qc/0607039 [gr-qc]]. 22

    2006 arXiv

  5. [13]

    Taming Thiemann’s 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 Hamiltonian constraint in canonical loop quantum gravity: reversibility, eigenstates and graph- change analysis,” [arXiv:2412.20272 [gr-qc]]

    arXiv

  6. [14]

    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,” [arXiv:2412.20257 [gr-qc]]

  7. [15]

    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]]

    2024 arXiv

  8. [16]

    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]]

    2024 arXiv

  9. [17]

    World sheet formulations of gauge theories and gravity,

    M. P. Reisenberger, “World sheet formulations of gauge theories and gravity,” [arXiv:gr-qc/9412035 [gr-qc]]

    arXiv

  10. [18]

    ’Sum over surfaces’ form of loop quantum gravity,

    M. P. Reisenberger and C. Rovelli, “’Sum over surfaces’ form of loop quantum gravity,” Phys. Rev. D 56 (1997), 3490-3508 doi:10.1103/PhysRevD.56.3490 [arXiv:gr-qc/9612035 [gr-qc]]

    1997 arXiv

  11. [19]

    Covariant Loop Quantum Gravity: An Elementary Introduction to Quantum Gravity and Spinfoam Theory,

    C. Rovelli and F. Vidotto, “Covariant Loop Quantum Gravity: An Elementary Introduction to Quantum Gravity and Spinfoam Theory,” Cambridge University Press, 2014, ISBN 978-1-107- 06962-6, 978-1-316-14729-0

    2014

  12. [20]

    The Spin Foam Approach to Quantum Gravity,

    A. Perez, “The Spin Foam Approach to Quantum Gravity,” Living Rev. Rel. 16 (2013), 3 doi:10.12942/lrr-2013-3 [arXiv:1205.2019 [gr-qc]]

    2013 arXiv

  13. [21]

    Spin foam models,

    J. C. Baez, “Spin foam models,” Class. Quant. Grav. 15 (1998), 1827-1858 doi:10.1088/0264- 9381/15/7/004 [arXiv:gr-qc/9709052 [gr-qc]]

    1998 arXiv

  14. [22]

    Handbook of Quantum Gravity,

    C. Bambi, L. Modesto and I. Shapiro, “Handbook of Quantum Gravity,” Springer, 2024, ISBN 978-981-99-7680-5, 978-981-99-7681-2, 978-981-19-3079-9 doi:10.1007/978-981-99-7681-2

    2024 doi

  15. [23]

    Spin Foams: Foundations,

    J. Engle and S. Speziale, “Spin Foams: Foundations,” doi:10.1007/978-981-19-3079-9 99-1 [arXiv:2310.20147 [gr-qc]]

    arXiv

  16. [24]

    Spinfoam Models for Quantum Gravity: Overview,

    E. R. Livine, “Spinfoam Models for Quantum Gravity: Overview,” doi:10.1016/B978-0-323-95703- 8.00253-6 [arXiv:2403.09364 [gr-qc]]

    arXiv

  17. [25]

    Spin-Foams for All Loop Quantum Gravity,

    W. Kaminski, M. Kisielowski and J. Lewandowski, “Spin-Foams for All Loop Quantum Gravity,” Class. Quant. Grav. 27 (2010), 095006 [erratum: Class. Quant. Grav. 29 (2012), 049502] doi:10.1088/0264-9381/29/4/049502 [arXiv:0909.0939 [gr-qc]]

    2010 arXiv

  18. [26]

    Operator Spin Foam Models,

    B. Bahr, F. Hellmann, W. Kaminski, M. Kisielowski and J. Lewandowski, “Operator Spin Foam Models,” Class. Quant. Grav. 28 (2011), 105003 doi:10.1088/0264-9381/28/10/105003 [arXiv:1010.4787 [gr-qc]]

    2011 arXiv

  19. [27]

    The Loop-quantum-gravity vertex-amplitude,

    J. Engle, R. Pereira and C. Rovelli, “The Loop-quantum-gravity vertex-amplitude,” Phys. Rev. Lett. 99 (2007), 161301 doi:10.1103/PhysRevLett.99.161301 [arXiv:0705.2388 [gr-qc]]

    2007 arXiv

  20. [28]

    Spinfoams and High-Performance Computing,

    P. Dona, M. Han and H. Liu, “Spinfoams and High-Performance Computing,” doi:10.1007/978- 981-19-3079-9 100-1 [arXiv:2212.14396 [gr-qc]]

    arXiv

  21. [29]

    Numerical study of the Lorentzian Engle-Pereira-Rovelli-Livine spin foam amplitude,

    P. Don` a, M. Fanizza, G. Sarno and S. Speziale, “Numerical study of the Lorentzian Engle-Pereira-Rovelli-Livine spin foam amplitude,” Phys. Rev. D 100 (2019) no.10, 106003 doi:10.1103/PhysRevD.100.106003 [arXiv:1903.12624 [gr-qc]]

    2019 arXiv

  22. [30]

    A high-performance code for EPRL spin foam amplitudes,

    F. Gozzini, “A high-performance code for EPRL spin foam amplitudes,” Class. Quant. Grav. 38 (2021) no.22, 225010 doi:10.1088/1361-6382/ac2b0b [arXiv:2107.13952 [gr-qc]]

    2021 arXiv

  23. [31]

    How-to Compute EPRL Spin Foam Amplitudes,

    P. Dona and P. Frisoni, “How-to Compute EPRL Spin Foam Amplitudes,” Universe 8 (2022) no.4, 208 doi:10.3390/universe8040208 [arXiv:2202.04360 [gr-qc]]

    2022 arXiv

  24. [32]

    LQG vertex with finite Immirzi parameter,

    J. Engle, E. Livine, R. Pereira and C. Rovelli, “LQG vertex with finite Immirzi parameter,” Nucl. Phys. B 799 (2008), 136-149 doi:10.1016/j.nuclphysb.2008.02.018 [arXiv:0711.0146 [gr-qc]]

    2008 arXiv

  25. [33]

    Numerical analysis of spin foam dynamics and the flatness problem,

    P. Don` a, F. Gozzini and G. Sarno, “Numerical analysis of spin foam dynamics and the flatness problem,” Phys. Rev. D 102 (2020) no.10, 106003 doi:10.1103/PhysRevD.102.106003 [arXiv:2004.12911 [gr-qc]]

    2020 arXiv

  26. [34]

    Radiative corrections to the Lorentzian Engle- 23 Pereira-Rovelli-Livine spin foam propagator,

    P. Don` a, P. Frisoni and E. Wilson-Ewing, “Radiative corrections to the Lorentzian Engle- 23 Pereira-Rovelli-Livine spin foam propagator,” Phys. Rev. D 106 (2022) no.6, 066022 doi:10.1103/PhysRevD.106.066022 [arXiv:2206.14755 [gr-qc]]

    2022 arXiv

  27. [35]

    Markov chain Monte Carlo methods for graph refinement in spinfoam cosmology,

    P. Frisoni, F. Gozzini and F. Vidotto, “Markov chain Monte Carlo methods for graph refinement in spinfoam cosmology,” Class. Quant. Grav. 40 (2023) no.10, 105001 doi:10.1088/1361-6382/acc5d6 [arXiv:2207.02881 [gr-qc]]

    2023 arXiv

  28. [36]

    Spinfoam on a Lefschetz thimble: Markov chain Monte Carlo computation of a Lorentzian spinfoam propagator,

    M. Han, Z. Huang, H. Liu, D. Qu and Y. Wan, “Spinfoam on a Lefschetz thimble: Markov chain Monte Carlo computation of a Lorentzian spinfoam propagator,” Phys. Rev. D 103 (2021) no.8, 084026 doi:10.1103/PhysRevD.103.084026 [arXiv:2012.11515 [gr-qc]]

    2021 arXiv

  29. [37]

    Summing bulk quantum numbers with Monte Carlo in spin foam theories,

    P. Don` a and P. Frisoni, “Summing bulk quantum numbers with Monte Carlo in spin foam theories,” Phys. Rev. D 107 (2023) no.10, 106008 doi:10.1103/PhysRevD.107.106008 [arXiv:2302.00072 [gr- qc]]

    2023 arXiv

  30. [38]

    Generative Flow Networks in Covariant Loop Quantum Gravity,

    J. Bunao, P. Frisoni, A. Kogios and J. Wogan, “Generative Flow Networks in Covariant Loop Quantum Gravity,” [arXiv:2407.19036 [gr-qc]]

    arXiv

  31. [39]

    Monte Carlo algorithm for spin foam intertwiners,

    S. Steinhaus, “Monte Carlo algorithm for spin foam intertwiners,” Phys. Rev. D 110 (2024) no.2, 026022 doi:10.1103/PhysRevD.110.026022 [arXiv:2403.04836 [gr-qc]]

    2024 arXiv

  32. [40]

    Efficient tensor network algorithms for spin foam models,

    S. K. Asante and S. Steinhaus, “Efficient tensor network algorithms for spin foam models,” Phys. Rev. D 110 (2024) no.10, 106018 doi:10.1103/PhysRevD.110.106018 [arXiv:2406.19676 [gr-qc]]

    2024 arXiv

  33. [41]

    Complex critical points and curved geometries in four-dimensional Lorentzian spinfoam quantum gravity,

    M. Han, Z. Huang, H. Liu and D. Qu, “Complex critical points and curved geometries in four-dimensional Lorentzian spinfoam quantum gravity,” Phys. Rev. D 106 (2022) no.4, 044005 doi:10.1103/PhysRevD.106.044005 [arXiv:2110.10670 [gr-qc]]

    2022 arXiv

  34. [42]

    Complex critical points in Lorentzian spinfoam quantum gravity: Four-simplex amplitude and effective dynamics on a double-∆3 complex,

    M. Han, H. Liu and D. Qu, “Complex critical points in Lorentzian spinfoam quantum gravity: Four-simplex amplitude and effective dynamics on a double-∆3 complex,” Phys. Rev. D 108 (2023) no.2, 026010 doi:10.1103/PhysRevD.108.026010 [arXiv:2301.02930 [gr-qc]]

    2023 arXiv

  35. [43]

    Mathematica program for numerically computing real and complex critical points in four-dimensional Lorentzian spinfoam amplitudes,

    M. Han, H. Liu and D. Qu, “Mathematica program for numerically computing real and complex critical points in four-dimensional Lorentzian spinfoam amplitudes,” Phys. Rev. D 111 (2025) no.2, 024021 doi:10.1103/PhysRevD.111.024021 [arXiv:2404.10563 [gr-qc]]

    2025 arXiv

  36. [44]

    Spin-foams as semiclassical vertices: Gluing constraints and a hybrid algorithm,

    S. K. Asante, J. D. Sim˜ ao and S. Steinhaus, “Spin-foams as semiclassical vertices: Gluing constraints and a hybrid algorithm,” Phys. Rev. D 107 (2023) no.4, 046002 doi:10.1103/PhysRevD.107.046002 [arXiv:2206.13540 [gr-qc]]

    2023 arXiv

  37. [45]

    Bahr and S

    B. Bahr and S. Steinhaus, Phys. Rev. D 93 (2016) no.10, 104029 doi:10.1103/PhysRevD.93.104029 [arXiv:1508.07961 [gr-qc]]

    2016 arXiv

  38. [46]

    Numerical evidence for a phase transition in 4d spin foam quantum gravity,

    B. Bahr and S. Steinhaus, “Numerical evidence for a phase transition in 4d spin foam quantum gravity,” Phys. Rev. Lett.117 (2016) no.14, 141302 doi:10.1103/PhysRevLett.117.141302 [arXiv:1605.07649 [gr-qc]]

    2016 arXiv

  39. [47]

    Effective Spin Foam Models for Four-Dimensional Quantum Gravity,

    S. K. Asante, B. Dittrich and H. M. Haggard, “Effective Spin Foam Models for Four-Dimensional Quantum Gravity,” Phys. Rev. Lett. 125 (2020) no.23, 231301 doi:10.1103/PhysRevLett.125.231301 [arXiv:2004.07013 [gr-qc]]

    2020 arXiv

  40. [48]

    Effective spin foam models for Lorentzian quantum gravity,

    S. K. Asante, B. Dittrich and J. Padua-Arguelles, “Effective spin foam models for Lorentzian quantum gravity,” Class. Quant. Grav. 38 (2021) no.19, 195002 doi:10.1088/1361-6382/ac1b44 [arXiv:2104.00485 [gr-qc]]

    2021 arXiv

  41. [49]

    Discrete gravity dynamics from effective spin foams,

    S. K. Asante, B. Dittrich and H. M. Haggard, “Discrete gravity dynamics from effective spin foams,” Class. Quant. Grav. 38 (2021) no.14, 145023 doi:10.1088/1361-6382/ac011b [arXiv:2011.14468 [gr-qc]]

    2021 arXiv

  42. [50]

    Modified graviton dynamics from spin foams: the area Regge action,

    B. Dittrich, “Modified graviton dynamics from spin foams: the area Regge action,” Eur. Phys. J. Plus 139 (2024) no.7, 651 doi:10.1140/epjp/s13360-024-05432-4 [arXiv:2105.10808 [gr-qc]]

    2024 arXiv

  43. [51]

    From spin foams to area metric dynamics to gravitons,

    B. Dittrich and A. Kogios, “From spin foams to area metric dynamics to gravitons,” Class. Quant. Grav. 40 (2023) no.9, 095011 doi:10.1088/1361-6382/acc5d9 [arXiv:2203.02409 [gr-qc]]

    2023 arXiv

  44. [52]

    Towards effective actions for the continuum limit of spin foams,

    J. N. Borissova and B. Dittrich, “Towards effective actions for the continuum limit of spin foams,” Class. Quant. Grav. 40 (2023) no.10, 105006 doi:10.1088/1361-6382/accbfb [arXiv:2207.03307 [gr- qc]]

    2023 arXiv

  45. [53]

    Relativistic spin networks and quantum gravity,

    J. W. Barrett and L. Crane, “Relativistic spin networks and quantum gravity,” J. Math. Phys. 39 24 (1998), 3296-3302 doi:10.1063/1.532254 [arXiv:gr-qc/9709028 [gr-qc]]

    1998 arXiv

  46. [54]

    A Lorentzian signature model for quantum general relativity,

    J. W. Barrett and L. Crane, “A Lorentzian signature model for quantum general relativity,” Class. Quant. Grav. 17 (2000), 3101-3118 doi:10.1088/0264-9381/17/16/302 [arXiv:gr-qc/9904025 [gr-qc]]

    2000 arXiv

  47. [55]

    An Introduction to Spin Foam Models of BF Theory and Quantum Gravity,

    J. C. Baez, “An Introduction to Spin Foam Models of BF Theory and Quantum Gravity,” Lect. Notes Phys. 543 (2000), 25-93 doi:10.1007/3-540-46552-9 2 [arXiv:gr-qc/9905087 [gr-qc]]

    2000 arXiv

  48. [56]

    On relativistic spin network vertices,

    M. P. Reisenberger, “On relativistic spin network vertices,” J. Math. Phys. 40 (1999), 2046-2054 doi:10.1063/1.532850 [arXiv:gr-qc/9809067 [gr-qc]]

    1999 arXiv

  49. [57]

    Barrett-Crane spin foam model from generalized BF type action for gravity,

    R. E. Livine and D. Oriti, “Barrett-Crane spin foam model from generalized BF type action for gravity,” Phys. Rev. D 65 (2002), 044025 doi:10.1103/PhysRevD.65.044025 [arXiv:gr-qc/0104043 [gr-qc]]

    2002 arXiv

  50. [58]

    The Complete LQG propagator. I. Difficulties with the Barrett-Crane vertex,

    E. Alesci and C. Rovelli, “The Complete LQG propagator. I. Difficulties with the Barrett-Crane vertex,” Phys. Rev. D 76 (2007), 104012 doi:10.1103/PhysRevD.76.104012 [arXiv:0708.0883 [gr- qc]]

    2007 arXiv

  51. [59]

    On the separation of Einsteinian substructures,

    J. F. Plebanski, “On the separation of Einsteinian substructures,” J. Math. Phys. 18 (1977), 2511-2520 doi:10.1063/1.523215

    1977 doi

  52. [60]

    A New Spin Foam Model for 4d Gravity,

    L. Freidel and K. Krasnov, “A New Spin Foam Model for 4d Gravity,” Class. Quant. Grav. 25 (2008), 125018 doi:10.1088/0264-9381/25/12/125018 [arXiv:0708.1595 [gr-qc]]

    2008 arXiv

  53. [61]

    Sub-leading asymptotic behaviour of area correlations in the Barrett-Crane model,

    J. D. Christensen, I. Khavkine, E. R. Livine and S. Speziale, “Sub-leading asymptotic behaviour of area correlations in the Barrett-Crane model,” Class. Quant. Grav. 27 (2010), 035012 doi:10.1088/0264-9381/27/3/035012 [arXiv:0908.4476 [gr-qc]]

    2010 arXiv

  54. [62]

    Positivity of spin foam amplitudes,

    J. C. Baez and J. D. Christensen, “Positivity of spin foam amplitudes,” Class. Quant. Grav. 19 (2002), 2291-2306 doi:10.1088/0264-9381/19/8/316 [arXiv:gr-qc/0110044 [gr-qc]]

    2002 arXiv

  55. [63]

    An Efficient algorithm for the Riemannian 10j symbols,

    J. D. Christensen and G. Egan, “An Efficient algorithm for the Riemannian 10j symbols,” Class. Quant. Grav. 19 (2002), 1185-1194 doi:10.1088/0264-9381/19/6/310 [arXiv:gr-qc/0110045 [gr-qc]]

    2002 arXiv

  56. [64]

    LLVM: A Compilation Framework for Lifelong Program Analysis and Transformation,

    C. Lattner and V. Adve, “LLVM: A Compilation Framework for Lifelong Program Analysis and Transformation,” in Proceedings of the 2004 International Symposium on Code Generation and Optimization (CGO) , San Jose, CA, USA, Mar 2004, pp. 75-88

    2004

  57. [65]

    Asymptotics of 10j symbols,

    J. C. Baez, J. D. Christensen and G. Egan, “Asymptotics of 10j symbols,” Class. Quant. Grav. 19 (2002), 6489 doi:10.1088/0264-9381/19/24/315 [arXiv:gr-qc/0208010 [gr-qc]]

    2002 arXiv

  58. [66]

    Spin foam models of Riemannian quantum gravity,

    J. C. Baez, J. D. Christensen, T. R. Halford and D. C. Tsang, “Spin foam models of Riemannian quantum gravity,” Class. Quant. Grav. 19 (2002), 4627-4648 doi:10.1088/0264-9381/19/18/301 [arXiv:gr-qc/0202017 [gr-qc]]

    2002 arXiv

  59. [67]

    Learning representations by back-propagating errors,

    D. E. Rumelhart, G. E. Hinton and R. J. Williams, “Learning representations by back-propagating errors,” Nature 323 (1986) no.6088, 533-536 doi:10.1038/323533a0

    1986 doi

  60. [68]

    Multilayer feedforward networks are universal approximators,

    K. Hornik, M. Stinchcombe and H. White, “Multilayer feedforward networks are universal approximators,” Neural Networks 2 (1989) no.5, 359-366 doi:10.1016/0893-6080(89)90020-8

    1989 doi

  61. [69]

    PyTorch: An Imperative Style, High-Performance Deep Learning Library,

    A. Paszke, S. Gross, F. Massa, A. Lerer, J. Bradbury, G. Chanan, T. Killeen, Z. Lin, N. Gimelshein and L. Antiga, et al. “PyTorch: An Imperative Style, High-Performance Deep Learning Library,” [arXiv:1912.01703 [cs.LG]]

    1912 arXiv

  62. [70]

    Optuna: A Next-generation Hyperparameter Optimization Framework,

    T. Akiba, S. Sano, T. Yanase, T. Ohta and M. Koyama, “Optuna: A Next-generation Hyperparameter Optimization Framework,” [arXiv:1907.10902 [cs.LG]]

    1907 arXiv

  63. [71]

    Decoupled Weight Decay Regularization,

    I. Loshchilov and F. Hutter, “Decoupled Weight Decay Regularization,” [arXiv:1711.05101 [cs.LG]]

    arXiv

  64. [72]

    Evaluation: from precision, recall and F-measure to ROC, informedness, markedness and correlation,

    D. M. W. Powers, “Evaluation: from precision, recall and F-measure to ROC, informedness, markedness and correlation,” (2020). [arXiv:2010.16061 [cs.LG]]

    2020 arXiv

  65. [73]

    Robust Estimation of a Location Parameter,

    P. J. Huber, “Robust Estimation of a Location Parameter,” Annals of Mathematical Statistics 35 (1964), 73-101. doi:10.1214/aoms/1177703732

    1964

  66. [74]

    Adaptive Mixtures of Local Experts,

    R. A. Jacobs, M. I. Jordan, S. J. Nowlan and G. E. Hinton, “Adaptive Mixtures of Local Experts,” Neural Computation 3 (1991), 79-87. doi:10.1162/neco.1991.3.1.79

    1991 doi

  67. [75]

    Outrageously Large Neural Networks: The Sparsely-Gated Mixture-of-Experts Layer,

    N. Shazeer, A. Mirhoseini, K. Maziarz, A. Davis, Q. Le, G. Hinton and J. Dean, “Outrageously Large Neural Networks: The Sparsely-Gated Mixture-of-Experts Layer,” International Conference 25 on Learning Representations (ICLR) 2017. [arXiv:1701.06538 [cs.LG]]

    2017 arXiv

  68. [76]

    Proximal Policy Optimization Algorithms,

    J. Schulman, F. Wolski, P. Dhariwal, A. Radford and O. Klimov, “Proximal Policy Optimization Algorithms,” [arXiv:1707.06347 [cs.LG]]

    arXiv

  69. [77]

    Learning Phrase Representations using RNN Encoder - Decoder for Statistical Machine Translation,

    K. Cho, B. van Merri¨ enboer, C. Gulcehre, D. Bahdanau, F. Bougares, H. Schwenk and Y. Bengio, “Learning Phrase Representations using RNN Encoder - Decoder for Statistical Machine Translation,” Proceedings of the 2014 Conference on Empirical Methods in Natural Language Process...

    2014 arXiv

  70. [78]

    The Graph Neural Network Model,

    F. Scarselli, M. Gori, A. C. Tsoi, M. Hagenbuchner and G. Monfardini, “The Graph Neural Network Model,” IEEE Transactions on Neural Networks 20 (2009), 61-80. doi:10.1109/TNN.2008.2005605

    2009

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.