Pith. sign in

REVIEW 5 major objections 6 minor 50 references

Siamese Neural Network for Label-Efficient Critical Phenomena Prediction in 3D Percolation Models

T0 review · 5 major / 6 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read A Siamese neural network trained on just 22 labeled probability points from non-critical regions locates 3D percolation thresholds and the critical exponent ν.

desk verdict Plausible label-efficient p_c detection for 3D percolation, but the abstract overclaims: nu is fixed not estimated, FCC transfer and representation results are missing from the body, and the MC baseline is off by several sigma. read the letter →

arxiv 2507.14159 v2 pith:NOYHIOPJ submitted 2025-07-05 cond-mat.dis-nn cs.LG

Shanshan Wang , Dian Xu , Jianmin Shen , Feng Gao , Wei Li , Weibing Deng This is my paper · ORCID
classification cond-mat.dis-nncs.LG
keywords Siameseneuralnetworkpercolationthresholdcriticalexponent3Dsiteandbondlabel-efficientlearninglargest-clustersizefinite-sizescaling
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

The paper tries to establish that phase transitions in three-dimensional percolation can be located with almost no labeled data: a Siamese neural network trained only on binary similarity labels for pairs drawn from the far non-critical regions [0,0.1] and [0.9,1] estimates the percolation threshold with percent-level accuracy and yields the critical exponent ν consistent with literature values. The network was never shown a configuration near the transition, yet its learned representation quantitatively matches the normalized largest-cluster size S_max/$L^{3}$, the standard finite-size order parameter. If true, this offers a practical route to criticality detection in settings where no order parameter is explicitly defined and labeled data are scarce.

What carries the argument

The central machinery is a Siamese neural network with two shared-weight fully connected branches that embed DFS-extracted largest-cluster configurations into a latent space, followed by a similarity evaluator trained with binary cross-entropy on positive and negative pairs. The load-bearing identity is that the learned embedding collapses onto the normalized largest-cluster size S_max/$L^{3}$, so the network's similarity score effectively measures the finite-size order parameter of percolation; the threshold is read as the crossing of average similarity curves for same-region versus cross-region pairs, and finite-size scaling extrapolates that crossing to the thermodynamic limit.

What would settle it

Train the SNN with the same 22 labeled probability points but with anchors placed inside and immediately below the critical region, and compare the crossing points to the high-precision thresholds 0.3116 and 0.2488; if the crossing shifts systematically with anchor by more than the reported error bars, the extracted threshold is an artifact of the labeling protocol rather than the true critical point.

Watch

Extended reading notes

Core claim

The paper reports that, for site and bond percolation on a three-dimensional simple cubic lattice, a Siamese network trained solely on configuration pairs labeled as same-region or different-region, using 22 labeled probability points taken entirely from [0,0.1]∪[0.9,1], recovers the percolation threshold from the crossing of the positive and negative similarity curves. Finite-size scaling extrapolation gives p_c ≈ 0.309–0.315 for site percolation and ≈ 0.248–0.253 for bond percolation, compared with literature values of 0.3116 and 0.2488, and data collapse of the network outputs gives ν ≈ 0.88, consistent with the known value 0.8765. The paper further claims that the learned embedding coincides with the normalized largest-cluster size S_max/$L^{3}$ at correlation r > 0.99, and that a network trained solely on simple cubic lattices identifies the phase transition in face-centered cubic lattices without retraining.

Load-bearing premise

The paper assumes that the point where the positive and negative similarity curves cross marks the true percolation threshold, independent of the chosen anchor and labeling interval, even though the network was trained only on labels from outside the critical region.

Editorial extensions

If this is right

  • Three-dimensional site and bond percolation thresholds on cubic lattices can be extracted to within about one percent using only 22 labeled probability points, none of them near the critical region.
  • The same network output, after data collapse, yields the correlation-length exponent ν ≈ 0.88, matching the literature value 0.8765 within statistical uncertainty.
  • The learned representation is, up to a correlation exceeding 0.99, the normalized largest-cluster size S_max/L^3, indicating that the network discovers the finite-size order parameter from similarity labels alone.
  • A network trained on simple cubic lattices transfers to face-centered cubic lattices without retraining, suggesting the learned statistic is not tied to one lattice geometry.
  • Extending the labeling interval toward the critical region does not significantly improve the threshold or exponent estimates, supporting the claim of label efficiency.

Reading between the lines

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

  • If the learned embedding really is S_max/L^3, the same architecture should work for any model whose transition is governed by a spanning or largest-cluster observable, including continuum percolation and correlated percolation variants; that is a testable prediction beyond the paper.
  • The method could serve as an order-parameter discovery tool: when no quantitative order parameter is known, the SNN embedding itself may be used as the scaling variable in a finite-size collapse.
  • The anchor-to-anchor spread in the reported thermodynamic-limit thresholds (about 0.006 for site and 0.005 for bond, larger than the quoted extrapolation errors) invites a systematic study of how the crossing-point estimator depends on anchor location; a monotone trend would indicate a systematic component.
  • Applying the identical 22-point protocol to two-dimensional percolation, where p_c and ν are known to high precision, would provide a cheap external calibration of whether the crossing rule is unbiased.
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

5 major / 6 minor

Summary. The paper trains a Siamese neural network (SNN) on pairwise similarity labels derived from 3D site and bond percolation configurations, using as input the largest connected cluster extracted by depth-first search. Labels are generated only from configurations in the non-critical intervals [0,0.1] and [0.9,1], giving 22 labeled probability points across five system sizes. The authors propose that the intersection of the SNN's positive and negative similarity curves marks the percolation threshold, and they extract p_c via finite-size scaling. They also present data-collapse plots for the SNN output and report the critical exponent ν=0.88. The central advertised results are percent-level estimates of p_c, ν consistent with literature within statistical uncertainty, and transfer from simple cubic to face-centered cubic lattices without retraining.

Significance. If fully substantiated, a label-efficient SNN that locates 3D percolation thresholds and estimates critical exponents from only 22 labeled probability points would be a useful addition to the ML-for-critical-phenomena toolkit. The paper has several strengths: it uses a large set of independent Monte Carlo configurations per probability point, shares weights between the two branches, explicitly studies the effect of extending the labeling interval via iteration, and reports anchor-based robustness tables. However, the current manuscript does not support the critical-exponent claim or the face-centered-cubic transfer claim, and the p_c extraction procedure is insufficiently formalized. The significance is therefore conditional on substantial additional analysis.

major comments (5)
  1. [Abstract; §4.3; Figs. 7-8; Table 4] The claim that the method “yields estimates of the critical exponent ν consistent with literature values within statistical uncertainty” is not supported by the reported analysis. In §4.3 the data collapse is performed by setting ν≈0.88 beforehand (“when the critical exponent is set to ν≈0.88, the data curves … collapse”), and Table 4 lists ν=0.88 and ν_it=0.88 for every anchor, with no uncertainty, no goodness-of-fit measure, and no search over ν. The exponent is therefore an input chosen by the authors, not an estimate produced by the SNN. To substantiate or retract the abstract claim, the authors should perform a data-collapse fit in which ν is a free parameter, report the fitted value with uncertainty for each anchor, and state a quantitative collapse criterion.
  2. [Abstract; Sections 2-5] The abstract's headline transfer claim — that a network trained solely on simple cubic lattices identifies the phase transition in face-centered cubic lattices without retraining — has no corresponding experimental section, figure, or table anywhere in the main text. No fcc model, fcc simulation, or fcc result is described in Sections 2-5. Either add a complete fcc subsection with simulation details and quantitative results, or remove the claim from the abstract and introduction, since it is currently unsupported.
  3. [§4.2; §4.3; Tables 1-2] The procedure for extracting p_c is not formally defined. The text states that p_c is identified as the intersection of the average similarity curves of positive and negative sample pairs, but no equation or algorithm specifies how the curves are averaged, how the intersection is computed, or how uncertainties are propagated from individual similarity outputs to the finite-size estimates. The anchor-dependent thermodynamic-limit results in Tables 1-2 spread from 0.3090 to 0.3149 (site) and from 0.2484 to 0.2530 (bond); these spreads are larger than the reported uncertainties and are comparable to the claimed percent-level accuracy. The monotone-crossing premise also lacks a derivation. The authors should formalize the estimator, provide an error budget, and either demonstrate that anchor dependence is within statistical error or introduce an anchor-independent estimator.
  4. [§4.1; Figs. 4-5] The Monte Carlo calibration itself deviates from the accepted literature values by more than the reported statistical errors. Figure 4(f) gives p_c=0.3146(14) for 3D site percolation, while the text quotes the standard value 0.3116; Figure 5(f) gives p_c=0.2513(7) for bond percolation, versus 0.2488. These discrepancies are roughly 1% and are many times the stated uncertainties, yet the text says the results “align well with theoretical predictions.” Because these same configurations feed the SNN, the source of the bias should be identified (for example, the sigmoid fitting form, the choice of 1/L extrapolation without correction-to-scaling terms, or insufficient system sizes) and the calibration repeated or the discrepancy discussed explicitly.
  5. [§4.2; Labeling strategy; §4.3] There is a circularity risk in the labeling scheme that should be addressed with a control experiment. Positive labels are assigned when two configurations come from the same interval in {[0,0.1], [0.9,1]} and negative labels when they come from different intervals. This target already places the two labeled classes on opposite sides of any plausible p_c, so the network is effectively trained to output a monotone function of p; reading off a threshold from that monotone function may reflect the label construction rather than the true transition. The paper should test this by training the identical SNN on labels derived from an arbitrary split of the same probabilities (for example, positive for p in [0,0.1] vs. negative for p in [0.45,0.55] on a model with no transition at 0.3), or by shuffling the p-values attached to configurations, and showing that the extracted intersection no longer tracks the true p_c. Such a control would directly address whether the representation and the intersection carry information beyond the prescribed label intervals.
minor comments (6)
  1. [§4.2] There is a duplicated text fragment: “Unlike traditional supervised learning that assigns Unlike traditional supervised learning, which assigns labels to individual samples…” should be corrected.
  2. [Abstract; §4.2] The phrase “22 labeled probability points” is ambiguous: each probability point contains 1000 configurations, so the labeled data volume is 22×1000 configurations per system size. Please state both the number of probability values and the number of configurations per value.
  3. [§4.1; Figs. 4-5] The statement that all sigmoid fits achieve “a goodness of fit exceeding 99.9%” is not defined; please report the specific metric (e.g., R², χ² per degree of freedom) and its value for each fit.
  4. [§3; Eqs. (4)-(6)] The architecture description is inconsistent: Eq. (4) says the similarity evaluator takes the concatenated embeddings, while Eq. (6) defines a distance D_W between embeddings and the text says that distance “or alternatively, the concatenated embeddings” is passed to the evaluator. Please specify which input the evaluator actually uses.
  5. [Captions of Figs. 6(b) and 6(e)] The captions read “FFS extrapolation”; this should be “FSS extrapolation” in both places.
  6. [Introduction; Conclusion] The phrase “first successful application of the SNN method for predicting critical thresholds in three-dimensional percolation models” is stronger than the evidence presented; earlier work cited in the paper (e.g., Ref. 24) already applies SNNs to phase diagrams, and no comparison with prior 3D SNN results is given. Please soften or support this claim.

Circularity Check

2 steps flagged · score 6.0 of 10

The advertised critical-exponent estimate is the input value ν≈0.88 reproduced in Table 4, and the claimed 'autonomous' learning of Smax/L^3 renames the DFS-extracted largest-cluster input.

  1. self definitional [Section 4.3 (data collapse) and Table 4; abstract]
    "when the critical exponent is set to ν ≈ 0.88, the data curves corresponding to different anchors in both the three-dimensional site and bond percolation models collapse onto a single universal curve, in good agreement with the theoretical value ν = 0.8765. Table 4 reports 'ν 0.88 0.88 0.88 0.88 0.88 0.8765' and 'νit 0.88 0.88 0.88 0.88'."

    The abstract claims the method 'yields estimates of the critical exponent ν consistent with literature values within statistical uncertainty,' but the data-collapse analysis does not estimate ν: it sets ν≈0.88 up front, then reports exactly that input value in Table 4 for every anchor, before and after iteration, with no uncertainty, no grid search, and no goodness-of-fit statistic. The reported 'estimate' is therefore the chosen parameter itself, reproduced verbatim; no statistical procedure connects the SNN curves to a fitted ν. The central ν claim reduces to the input by construction.

  2. renaming known result [Abstract and Section 4.2 (Input Data and Preprocessing for SNN)]
    "Analysis of the learned representations clarifies what the network learns: although trained solely on binary similarity labels, the network autonomously converges to a statistic that coincides quantitatively with the normalized largest-cluster size Smax/L^3 (r > 0.99), the finite-size order parameter of percolation. ... The largest cluster from each configuration was extracted using the Depth-First Search (DFS) algorithm. These extracted configurations were subsequently used as input pairs."

    The network input is already the DFS-extracted largest-cluster configuration; Smax/L^3 is a simple normalized sum of that binary tensor. Presenting the learned representation's quantitative agreement with Smax/L^3 as an 'autonomous convergence' renames the input statistic as a discovered order parameter. Since the input literally is the largest cluster, a representation reflecting its size is expected by construction: the network only needs to learn a weighted sum, so the claimed discovery is already contained in the feature supplied to the network.

full rationale

The p_c extraction is not circular: the 22 labels only mark far-below and far-above reference intervals, and locating the crossing of the similarity curves is an emergent discriminant, although the monotone-crossing premise is unproven and the anchor dependence in Tables 1 and 2 shows residual spread. The genuine circularity is confined to two load-bearing claims. First, the abstract advertises critical-exponent estimation, but Section 4.3 fixes ν≈0.88 as an input and Table 4 reports that same value; no fit, grid search, or uncertainty is provided, so the 'estimate' equals the assumption by construction. Second, the claim that the network 'autonomously converges' to Smax/L^3 is a renaming, because the input is already the DFS-extracted largest cluster and Smax/L^3 is its normalized sum. Self-citations (refs. 27 and 28) are only inspirational and do not carry the argument. Because the p_c prediction retains independent empirical content, the circularity is partial rather than total, giving 6/10.

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

The central outputs depend on hand-chosen anchors, hand-chosen labeling intervals, the untested similarity-crossing premise, and the largest-cluster feature choice. No code or independent experiment is provided to break these dependencies.

free parameters (4)
  • Anchor probability p_a = site: pa = 0, 0.15, 0.47, 1; bond: pa = 0, 0.12, 0.38, 1
    The threshold estimates in Tables 1 and 2 shift with anchor: site p_infinity_c ranges from 0.3090(20) to 0.3149(20), and bond from 0.2484(14) to 0.2530(19).
  • Labeling interval boundaries = 0.1 and 0.9; iterative refinement uses 0.99 and 0.01 similarity thresholds
    Only configurations with p in [0,0.1] union [0.9,1] are labeled. This hand-chosen split defines the 22 training probability points and the entire few-shot claim.
  • Critical exponent nu in data collapse = 0.88
    Nu is fixed at 0.88 in Figures 7-8 to show collapse, then reported in Table 4 with no uncertainty and no fitting procedure.
  • SNN training hyperparameters
    Batch size, epochs, learning rate, optimizer, random seeds, and number of runs are absent, leaving the exact trained model underspecified.
assumptions (4)
  • domain assumption Finite-size scaling relation Eq. (2) holds for the generated configurations with negligible corrections to scaling.
    Used both for MC validation and for the SNN data collapse; corrections to scaling are not assessed.
  • domain assumption The DFS-extracted largest-cluster representation is a sufficient input for locating the transition.
    The network never sees the full configuration; no test shows that full configurations would change the results.
  • ad hoc to paper The intersection of positive and negative similarity curves marks the true p_c.
    This is the testing protocol in Section 4.2. Tables 1-2 show the inferred p_c depends on the anchor, so the premise is only approximately true.
  • domain assumption A single critical exponent nu collapses the SNN output across system sizes and anchors.
    Figures 7-8 assume this universality and then visually inspect collapse with nu fixed at 0.88.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Siamese Neural Network for Label-Efficient Critical Phenomena Prediction in 3D Percolation Models." pith.science (2026). https://pith.science/paper/NOYHIOPJ

@misc{pith2026250714159,
  author       = {Pith},
  title        = {Pith review of: Siamese Neural Network for Label-Efficient Critical Phenomena Prediction in 3D Percolation Models},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/NOYHIOPJ}},
  note         = {Machine review of arXiv:2507.14159}
}
abstract

Predicting critical phenomena from limited labeled data remains a challenging task in statistical physics. As percolation theory provides a canonical model for phase transitions with well-established critical exponents, it serves as an ideal benchmark for validating new machine learning frameworks. Here, we introduce a label-efficient learning framework based on a Siamese Neural Network (SNN) to identify phase transitions in three-dimensional site and bond percolation models. Using only 22 labeled probability points drawn entirely from non-critical regions, the method locates percolation thresholds with percent-level accuracy and yields estimates of the critical exponent $\nu$ consistent with literature values within statistical uncertainty. Analysis of the learned representations clarifies what the network learns: although trained solely on binary similarity labels, the network autonomously converges to a statistic that coincides quantitatively with the normalized largest-cluster size $S_{max}/L^3$ ($r > 0.99$), the finite-size order parameter of percolation. This underlies the framework's most distinctive capability -- a model trained solely on simple cubic lattices identifies the phase transition in face-centered cubic lattices without retraining. The framework thus offers a complementary route to criticality detection in settings where no quantitative order parameter is explicitly defined or labeled data is scarce.

Figures

Figures reproduced from arXiv: 2507.14159 by the authors.

Figure 1
Figure 1. Site percolation on a three-dimensional cubic lattice. Top: Raw configurations at occupation probabilities p = 0.15, p = 0.312 (near the critical threshold pc ≈ 0.3116 for this system size), and p = 0.47. Bottom: Corresponding largest connected clusters, illustrating the transition from isolated clusters to a macroscopic spanning structure as p increases. The system size is L = 20. Occupied sites are shown; vacant s… view at source ↗
Figure 2
Figure 2. Bond percolation on a three-dimensional cubic lattice. Top: Raw configurations at occupation probabilities p = 0.12, p = 0.249 (near the critical threshold pc ≈ 0.2488 for this system size), and p = 0.38. Bottom: Corresponding largest connected clusters, illustrating the emergence of a macroscopic spanning structure as p increases. The system size is L = 10. Occupied (“open”) bonds are shown in red; vacant (“closed”… view at source ↗
Figure 3
Figure 3. Schematic Diagram of the SNN Architecture. weight matrix W, to map both inputs into a common latent representation space. A simplified form of the mapping can be expressed as: SW (x) = σ(Wx+b), (5) where σ is the activation function (e.g., Swish or Sigmoid), and b is the bias term. The distance between the latent representations SW (x1) and SW (x2) is then computed as: DW (x1, x2) = ∥SW (x1)−SW (x2)∥. (6) This dista… view at source ↗
Figures from the paper (6 more)
Figure 4
Figure 4. Figure 4: Monte Carlo simulation results for the 3D site percolation model. Panels (a)–(e) show results for system sizes L = 10, 16, 20, 24, and 30, respectively. The estimated critical threshold pc and its associated uncertainty are indicated in the upper-left corner of each su…
Figure 5
Figure 5. Figure 5: Monte Carlo simulation results for the 3D bond percolation model. Panels (a)–(e) show results for system sizes L = 10, 16, 20, 24, and 30, respectively. The estimated critical threshold pc and its associated uncertainty are indicated in the upper-left corner of each su…
Figure 6
Figure 6. Figure 6: The first row shows results for the three-dimensional site percolation model: (a) Similarity (dark) and dissimilarity (light) curves for various system sizes at anchor pa = 0.47; (b) FFS extrapolation for anchor pa = 0.47;(c) Output layer responses for various system s…
Figure 7
Figure 7. Figure 7: Data collapse analysis for estimating the critical exponent ν in the 3D site percolation model. Subfigures (a)–(d) correspond to anchor probabilities pa = 0, 0.15, 0.47, and 1, respectively. ◦◦◦◦◦◦◦◦◦◦ ◦◦◦◦◦◦◦◦ ◦ ◦ ◦ ◦ ◦ ◦ ◦ ◦ ◦◦◦ ◦ ◦◦◦◦◦◦◦◦◦◦◦◦◦◦◦◦◦◦◦◦ ◦◦◦◦◦◦◦◦◦◦ ◦◦◦◦…
Figure 8
Figure 8. Figure 8: Data collapse analysis for estimating the critical exponent ν in the 3D bond percolation model. Subfigures (a)–(d) correspond to anchor probabilities pa = 0, 0.12, 0.38, and 1, respectively. 9/14 [PITH_FULL_IMAGE:figures/full_fig_p009_8.png]
Figure 9
Figure 9. Figure 9: (a) Estimated critical thresholds at anchor pa = 0.47 for the three-dimensional site percolation model with system size L = 24, obtained via iterative adjustment of the probability labeling range. (b) FSS extrapolation at anchor pa = 0.47 as a function of 1/L, based on…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

50 extracted references · 48 canonical work pages

  1. [1]

    & Kumar, A

    Alzubi, J., Nayyar, A. & Kumar, A. Machine learning from theory to algorithms: an overview. In Journal of physics: conference series, vol. 1142, 012012 (IOP Publishing, 2018)

    work page 2018

  2. [2]

    & Hey, T

    Thiyagalingam, J., Shankar, M., Fox, G. & Hey, T. Scientific machine learning benchmarks. Nat. Rev. Phys. 4, 413–420 (2022)

    work page 2022

  3. [3]

    & Shashua, A

    Shalev-Shwartz, S., Shammah, S. & Shashua, A. Safe, multi-agent, reinforcement learning for autonomous driving. Arxiv Prepr. Arxiv:1610.03295 (2016)

    arXiv 2016

  4. [4]

    S., Shah, N

    Char, D. S., Shah, N. H. & Magnus, D. Implementing machine learning in health care—addressing ethical challenges. The New Engl. J. Medicine 378, 981 (2018)

    work page 2018

  5. [5]

    Y ., Hu, Y

    Lin, W. Y ., Hu, Y . H. & Tsai, C. F. Machine learning in financial crisis prediction: a survey.IEEE Transactions on Syst. Man, Cybern. Part C (Applications Rev. 42, 421–436 (2011)

    work page 2011

  6. [6]

    Jordan, M. I. & Mitchell, T. M. Machine learning: Trends, perspectives, and prospects. Science 349, 255–260 (2015)

    work page 2015

  7. [7]

    & Talwalkar, A

    Mohri, M., Rostamizadeh, A. & Talwalkar, A. Foundations of machine learning (The Mit Press, 2018)

    work page 2018

  8. [8]

    & Han, Z

    Xue, T., Li, X., Chen, X., Chen, L. & Han, Z. Machine learning phases in swarming systems. Mach. Learn. Sci. Technol. 4, 015028 (2023)

    work page 2023

Show all 50 references
  1. [9]

    Hu, G. et al. Universality class of machine learning for critical phenomena. Sci. China Physics, Mech. & Astron. 66, 120511 (2023)

    2023

  2. [10]

    Yang, Y . X.et al. Machine learning applications in phase transitions. Sci. Sinica Phys. Mech. & Astron. 53, 290011 (2023)

    2023

  3. [11]

    Lifshitz, E. M. & Pitaevskii, L. P. Statistical physics: theory of the condensed state, vol. 9 (Elsevier, 2013)

    2013

  4. [12]

    Monte Carlo: concepts, algorithms, and applications (Springer Science & Business Media, 2013)

    Fishman, G. Monte Carlo: concepts, algorithms, and applications (Springer Science & Business Media, 2013)

    2013

  5. [13]

    Newman, M. E. J. & Ziff, R. M. Efficient monte carlo algorithm and high-precision results for percolation. Phys. Rev. Lett. 85, 4104 (2000)

    2000

  6. [14]

    Hu, W., Singh, R. R. & Scalettar, R. T. Discovering phases, phase transitions, and crossovers through unsupervised machine learning: A critical examination. Phys. Rev. E 95, 062122 (2017)

    2017

  7. [15]

    P., Liu, Y .-H

    Van Nieuwenburg, E. P., Liu, Y .-H. & Huber, S. D. Learning phase transitions by confusion. Nat. Phys. 13, 435–439 (2017)

    2017

  8. [16]

    & Melko, R

    Carrasquilla, J. & Melko, R. G. Machine learning phases of matter. Nat. Phys. 13, 431–434 (2017)

    2017

  9. [17]

    Essam, J. W. Percolation theory. Reports on progress physics 43, 833 (1980)

    1980

  10. [18]

    & Seager, C

    Pike, G. & Seager, C. Percolation and conductivity: A computer study. i. Phys. review B 10, 1421 (1974)

    1974

  11. [19]

    & Ziff, R

    Newman, M. & Ziff, R. M. Efficient monte carlo algorithm and high-precision results for percolation. Phys. Rev. Lett. 85, 4104 (2000)

    2000

  12. [20]

    & Zhou, K

    Ma, Y .-G., Pang, L.-G., Wang, R. & Zhou, K. Phase transition study meets machine learning.Chin. Phys. Lett. 40, 122101 (2023). 12/14

    2023

  13. [21]

    Mehta, P. et al. A high-bias, low-variance introduction to machine learning for physicists. Phys. Reports 810, 1–124 (2019)

    2019

  14. [22]

    & Wei, T

    Zhang, W., Liu, J. & Wei, T. C. Machine learning of phase transitions in the percolation and xy models. Phys. Rev. E 99, 032142 (2019)

    2019

  15. [23]

    Li, X. et al. Machine learning phase transitions of the three-dimensional ising universality class. Chin. Phys. C 47, 034101 (2023)

    2023

  16. [24]

    & Wetzel, S

    Patel, Z., Merali, E. & Wetzel, S. J. Unsupervised learning of rydberg atom array phase diagram with siamese neural networks. New J. Phys. 24, 113021 (2022)

    2022

  17. [25]

    Discovering phase transitions with unsupervised learning

    Wang, L. Discovering phase transitions with unsupervised learning. Phys. Rev. B 94, 195105 (2016)

    2016

  18. [26]

    Wetzel, S. J. Unsupervised learning of phase transitions: From principal component analysis to variational autoencoders. Phys. Rev. E 96, 022140 (2017)

    2017

  19. [27]

    Shen, J. M. et al. Transfer learning of phase transitions in percolation and directed percolation. Phys. Rev. E 105, 064139 (2022)

    2022

  20. [28]

    Chen, X. N. et al. Study of phase transition of potts model with domain adversarial neural network. Phys. A: Stat. Mech. its Appl. 617, 128666 (2023)

    2023

  21. [29]

    Chen, X. N. et al. Applications of domain adversarial neural network in phase transition of 3d potts model. Phys. A: Stat. Mech. its Appl. 637, 129533 (2024)

    2024

  22. [30]

    Koch, G., Zemel, R., Salakhutdinov, R. et al. Siamese neural networks for one-shot image recognition. In ICML deep learning workshop, vol. 2, 1–30 (Lille, 2015)

    2015

  23. [31]

    & Aharony, A

    Stauffer, D. & Aharony, A. Introduction to percolation theory (Taylor & Francis, 2018)

    2018

  24. [32]

    Shante, V . K. & Kirkpatrick, S. An introduction to percolation theory.Adv. Phys. 20, 325–357 (1971)

    1971

  25. [33]

    Saberi, A. A. Recent advances in percolation theory and its applications. Phys. Reports 578, 1–32 (2015)

    2015

  26. [34]

    Isichenko, M. B. Percolation, statistical topography, and transport in random media. Rev. modern physics 64, 961 (1992)

    1992

  27. [35]

    & Ziff, R

    Araújo, N., Grassberger, P., Kahng, B., Schrenk, K. & Ziff, R. M. Recent advances and open challenges in percolation. The Eur. Phys. J. Special Top. 223, 2307–2321 (2014)

    2014

  28. [36]

    & Moloney, N

    Christensen, K. & Moloney, N. R. Complexity and criticality, vol. 1 (World Scientific Publishing Company, 2005)

    2005

  29. [37]

    Fessler, H. E. & Macklem, P. T. Percolation and phase transitions. Am. J. Respir. Critical Care Medicine 176, 530–531 (2007)

    2007

  30. [38]

    & Mitkov, R

    Ranasinghe, T., Orˇasan, C. & Mitkov, R. Semantic textual similarity with siamese neural networks. In Proceedings of the international conference on recent advances in natural language processing (RANLP 2019), 1004–1011 (2019)

    2019

  31. [39]

    Siamese neural networks: An overview

    Chicco, D. Siamese neural networks: An overview. Artif. neural networks 73–94 (2021)

    2021

  32. [40]

    Zhang, C., Liu, W., Ma, H. & Fu, H. Siamese neural network based gait recognition for human identification. In 2016 ieee international conference on acoustics, speech and signal processing (ICASSP), 2832–2836 (IEEE, 2016)

    2016

  33. [41]

    & Wolf, L

    Taigman, Y ., Yang, M., Ranzato, M. & Wolf, L. Deepface: Closing the gap to human-level performance in face verification. In Proceedings of the IEEE conference on computer vision and pattern recognition, 1701–1708 (2014)

    2014

  34. [42]

    Manocha, P. et al. Content-based representations of audio using siamese neural networks. In 2018 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 3136–3140 (IEEE, 2018)

    2018

  35. [43]

    & Blöte, H

    Deng, Y . & Blöte, H. W. Monte carlo study of the site-percolation model in two and three dimensions.Phys. Rev. E 72, 016126 (2005)

    2005

  36. [44]

    & Grossman, J

    Xie, T. & Grossman, J. C. Crystal graph convolutional neural networks for an accurate and interpretable prediction of material properties. Phys. review letters 120, 145301 (2018)

    2018

  37. [45]

    Peixoto, T. P. Bayesian stochastic blockmodeling. Adv. network clustering blockmodeling 289–332 (2019)

    2019

  38. [46]

    & Zhang, D

    Wang, N., Chang, H. & Zhang, D. Deep-learning-based inverse modeling approaches: A subsurface flow example. J. Geophys. Res. Solid Earth 126, e2020JB020549 (2021)

    2021

  39. [47]

    & Ringel, Z

    Koch-Janusz, M. & Ringel, Z. Mutual information, neural networks and the renormalization group. Nat. Phys. 14, 578–582 (2018)

    2018

  40. [48]

    & Wang, L

    Li, S.-H. & Wang, L. Neural network renormalization group. Phys. review letters 121, 260601 (2018). 13/14

    2018

  41. [49]

    Non-equilibrium phase transitions (Springer, 2008)

    Henkel, M. Non-equilibrium phase transitions (Springer, 2008)

    2008

  42. [50]

    Täuber, U. C. Critical dynamics: a field theory approach to equilibrium and non-equilibrium scaling behavior (Cambridge University Press, 2014). 14/14

    2014

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.