Pith. sign in

REVIEW 3 major objections 5 minor 55 references

Data-Driven Approach to Model the Influence of Magnetic Geometry in the Confinement of Fusion Devices

T0 review · 3 major / 5 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 Across a 12-million-configuration stellarator database, four boundary Fourier coefficients dominate omnigenity while two are nearly irrelevant.

desk verdict Big new VMEC database and a robust empirical least-important finding, but the importance ranking is distribution-dependent and unverified near omnigenity. read the letter →

arxiv 2507.03776 v1 pith:RXX37W5V submitted 2025-07-04 physics.plasm-ph

R. Laia , R. Jorge , G. Abreu This is my paper · ORCID
classification physics.plasm-ph PACS 52.55.Hc
keywords stellaratoromnigenityquasisymmetryquasi-isodynamicityboundaryFouriercoefficientsmachinelearningsurrogatesfeatureimportanceidealMHDequilibria
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 asks which geometric features of a stellarator's plasma boundary actually control whether the device will confine fusion-born $\alpha$ particles. To answer it, the authors generate a database of about 12.4 million vacuum stellarator equilibria, each described by eight boundary Fourier coefficients, and train several machine-learning models to predict two standard omnigenity metrics: quasisymmetry and quasi-isodynamicity. Their central finding is a ranking of the eight coefficients: the $m=1,n=0$ terms $RBC_{1,0}$ and $ZBS_{1,0}$ are consistently the least important for both metrics, while the $m=1,n=\pm 1$ terms $RBC_{\pm 1,1}$ and $ZBS_{\pm 1,1}$ dominate, according to gradient-boosting feature importance, attribution values, and mutual information. The best surrogate, a feed-forward neural network, predicts the logarithm of both omnigenity metrics on held-out data with $R^2=0.9909$. If the ranking holds in realistic designs, it tells stellarator optimizers which boundary degrees of freedom deserve the most attention and gives them cheap surrogate models to speed up the search.

What carries the argument

The load-bearing object is the Fourier representation of the plasma boundary, $R(\theta,\phi)=\sum RBC_{m,n}\cos(m\theta-n_{fp}n\phi)$ and $Z(\theta,\phi)=\sum ZBS_{m,n}\sin(m\theta-n_{fp}n\phi)$, truncated at $m_{pol}=n_{tor}=1$, which reduces the search space to eight coefficients with two field periods and major radius $R_0=1$. Each sampled boundary is fed to an ideal-MHD equilibrium solver that produces the magnetic field, from which the omnigenity proxies $f_{QS}$ and $f_{QI}$ and auxiliary quantities (rotational transform, mirror ratio, elongation, magnetic well, shear, inverse aspect ratio) are computed. The argument that some coefficients matter more than others is carried by three independent feature-attribution tools—gradient-boosting split counts, game-theoretic attribution values, and mutual information—plus a supervised autoencoder whose two-dimensional latent space displays a gradient of both metrics along one axis.

What would settle it

Run the same sampling procedure with $m_{pol}=n_{tor}=2$ (or with finite plasma pressure) and recompute the feature rankings; if $RBC_{1,0}$ or $ZBS_{1,0}$ rises out of the bottom two positions, the central ranking fails to transfer. A cheaper check is to take a published precisely omnigenous equilibrium, perturb $RBC_{1,0}$ and $ZBS_{1,0}$ by an amplitude comparable to perturbations of the claimed dominant coefficients, and measure whether $f_{QS}$, $f_{QI}$, or the radial drift degrades as much.

Watch

Extended reading notes

Core claim

The discovery the paper is trying to establish is that, within the eight-parameter family of two-field-period vacuum stellarators studied, omnigenity is governed by cross-section shaping coefficients with $|n|=1$ rather than by the two $n=0$ size-like coefficients. Concretely, the boundary Fourier coefficients $RBC_{1,1}$, $RBC_{-1,1}$, $ZBS_{1,1}$, and $ZBS_{-1,1}$ carry most of the predictive signal for both quasisymmetry and quasi-isodynamicity, while $RBC_{1,0}$ and $ZBS_{1,0}$ sit at the bottom of every importance ranking the paper computes. The paper further claims that these rankings are robust across method families—split-count importance in gradient-boosted trees, game-theoretic attribution, and mutual information—and that a feed-forward network trained on the same eight inputs reproduces both targets with $R^2=0.9909$. It concludes that the resulting models can act as fast surrogates for the ideal-MHD solver and that the supervised autoencoder's two-dimensional latent space offers a reduced design space aligned with omnigenity.

Load-bearing premise

The database only contains vacuum equilibria with two field periods, major radius one meter, and eight boundary Fourier coefficients drawn uniformly from $[-0.1,0.1]$, and the paper assumes this restricted family is representative enough that the feature-importance ranking carries over to realistic stellarators with many more modes and finite pressure.

Editorial extensions

If this is right

  • Stellarator optimizers in the same geometry family can concentrate search effort on the four $|n|=1$ coefficients, potentially reducing the effective dimension of the design problem from eight to four.
  • The trained feed-forward network and gradient-boosting regressors provide predictions of $\log f_{QS}$ and $\log f_{QI}$ accurate to $R^2 \approx 0.99$ in this family, so they can screen candidate configurations without running the expensive ideal-MHD solver.
  • The classification model, with about 99% accuracy on the test set, can pre-filter boundaries that will either fail to converge or fail to reach quasisymmetry below a residual of 10.
  • Quasisymmetry and mirror ratio are strongly correlated in this database, so mirror ratio is a useful cheap proxy during early design.
  • The supervised autoencoder reduces the eight coefficients to two latent coordinates in which omnigenity varies monotonically along one axis, suggesting optimization could be performed in that two-dimensional space.

Reading between the lines

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

  • The least-importance of $RBC_{1,0}$ and $ZBS_{1,0}$ may partly be a scale artifact: in a high-aspect-ratio, vacuum, fixed-$R_0$ family these coefficients mostly set the overall cross-section size, which cancels in the dimensionless omnigenity proxies; at lower aspect ratio or finite plasma pressure, size-dependent equilibrium shifts could make them matter more.
  • A direct test of transferability would be to repeat the sampling with $m_{pol}=n_{tor}=2$ or with finite beta; if the same two coefficients stay at the bottom, the ranking is a property of omnigenity itself rather than of the restricted boundary representation.
  • Because the surrogate models are trained on the paper's random sample, they are most reliable where the sample is dense; using them as optimizers would require an active-learning loop that adds new ideal-MHD evaluations in underrepresented low-residual regions, where the reported errors were largest.
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 / 5 minor

Summary. The paper builds a large database of 12,421,004 VMEC vacuum stellarator equilibria, all with nfp=2, mpol=ntor=1, and R0=1, where the eight nonzero boundary Fourier coefficients are sampled uniformly in [-0.1, 0.1]. For each equilibrium the authors compute the quasisymmetry proxy f_QS (Eq. 5), the quasi-isodynamicity proxy f_QI (Eq. 6), and several auxiliary physics metrics. They then perform correlation analysis, outlier detection, PCA, and a supervised autoencoder, and train LightGBM, LightGBM LSS, and feed-forward neural network regressors to predict log f_QS and log f_QI. Feature-importance analyses (native gain, SHAP, and mutual information) are used to rank the eight boundary coefficients. The central claim is that RBC1,0 and ZBS1,0 are consistently the least important coefficients for omnigenity, while the m=1, n=±1 coefficients dominate, and that the FFNN predicts log f_QS and log f_QI on held-out data with R²=0.9909.

Significance. If the feature-importance result is robust, the paper would provide a useful, openly available database and fast surrogate models for stellarator design-space exploration, plus a concrete data-driven statement about which low-order boundary modes matter for quasisymmetry and quasi-isodynamicity. The strengths are the open database and code, the use of VMEC-generated targets (so the surrogate training is not circular), and the consistency of the least-important-pair finding across several independent methods. However, the significance is capped by the deliberately narrow 8-parameter, vacuum, nfp=2 boundary family and by the fact that the importance rankings are computed over a population dominated by far-from-omnigenous configurations; the paper's own text acknowledges that precise omnigenity requires additional boundary modes. The claimed insight therefore currently holds only for the sampled distribution, not for the near-omnigenous regime where the proposed design conclusions and optimization surrogates would be used.

major comments (3)
  1. [Section VII and Figs. 13-17] The load-bearing conclusion that RBC1,0 and ZBS1,0 are the least important coefficients for omnigenity is drawn from importance measures computed on the full database, which is dominated by poor configurations: only 457,488 of 12,421,004 configurations have f_QS below 10, and in the classification test set only 91,500 of 2.49 million instances satisfy the converged-and-f_QS<10 condition (Section II and Table III). Tree-gain importance, SHAP values, and mutual information are distribution-dependent summaries: they describe how inputs move predictions over the sampled uniform prior, not which inputs control the approach to omnigenity in the low-residual region where optimization actually operates. To support the stated goal of quantifying the role of each surface degree of freedom in the loss functions used to find precisely omnigenous designs, the authors should recompute the importance rankings on a near-omnigenous subset (for example f_QS<10 and a corresponding f_QI threshold) and, if possible, compare with feature sensitivities for precisely omnigenous equilibria from the literature, such as Landreman-Paul quasisymmetric and Goodman et al. quasi-isodynamic configurations.
  2. [Section VI, Figs. 16-17] The FFNN SHAP analysis is performed with Kernel SHAP using a single-sample background drawn from the training data. With only one background sample, the conditional expectations required for Shapley-value computation are very crudely approximated, especially for eight correlated inputs, so the quantitative claim that RBC1,0 and ZBS1,0 have 'little to no impact' is not reliably established by this particular analysis. The authors should either use a sufficiently large background set (typically hundreds of samples), or corroborate the FFNN importance with an additional attribution method such as integrated gradients or permutation importance.
  3. [Section II, Eq. (4)] The database restricts the boundary to mpol=ntor=1 with nfp=2, and the text notes that precise omnigenity requires more Fourier modes. Consequently, the conclusion that RBC1,0 and ZBS1,0 are unimportant, and even the surrogate models themselves, are statements about this 8-parameter family rather than about magnetic geometry in general. The abstract and conclusions should carry this qualification explicitly, and the paper should either demonstrate that the rankings persist in an extended family or discuss how the omitted higher-order modes could couple to the modes studied here.
minor comments (5)
  1. [Abstract and Section I] There are typos in the omnigenity terminology: 'quasi-isodynamiticity' in the abstract and 'quasi-isodinamicity' in the introduction should be 'quasi-isodynamicity'.
  2. [Section V, Table III] The text states that the classification model 'predicted VMEC convergence and quasisymmetry below 10, predicting 99.7% of the configurations with those conditions correctly,' but Table III reports recall of only 0.85 for label 1; 99.7% appears to refer to overall accuracy. This should be stated unambiguously.
  3. [Section V, Figs. 13-14] The quantitative feature rankings differ between LightGBM, LightGBM LSS, and the FFNN SHAP results; only the least-important pair RBC1,0 and ZBS1,0 is consistent across methods. The text should explicitly acknowledge this disagreement and avoid implying that the models agree on the ordering of the dominant coefficients.
  4. [Section III, Fig. 5] The statement that quasisymmetry values 'can span up to 100 different orders of magnitude' should be clarified, since the caption of Table I lists quasisymmetry in [0,10] while the text discusses log-transformed values; the reader should know whether the reported range refers to the raw residual or its logarithm.
  5. [Section VI] The description of the FFNN architectures for the quasisymmetry and quasi-isodynamic models is identical, and both are reported to achieve R²=0.9909; the authors should confirm that these are indeed two distinct networks and report their train/validation/test splits consistently.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the targets are independent VMEC physics outputs, the surrogate models are validated on held-out data, and the feature-importance conclusions are empirical interpretations rather than inputs renamed as predictions.

full rationale

The paper's claimed derivation chain is data-driven rather than first-principles: the database Fourier coefficients x are sampled, VMEC solves the ideal MHD equilibrium, and the targets f_QS and f_QI are computed from the VMEC magnetic fields via Eqs. (5) and (6). Those targets are therefore physics outputs independent of the later machine-learning models. The LightGBM, LightGBM LSS, and feed-forward neural network models are trained to reproduce those targets and are evaluated on held-out configurations, so the reported R^2=0.9909 is a genuine out-of-sample predictive check, not a fitted quantity being relabeled as a prediction. The claims that RBC1,0 and ZBS1,0 are least important are drawn from tree-split gain, SHAP values, and mutual information computed on fitted models; these are interpretability summaries of the learned mapping over the sampled distribution, not a derivation that equates the conclusion with the inputs by construction. The supervised autoencoder's latent-space gradient is partly engineered by the regression term in Eq. (8), but the paper presents that as a representation-learning visualization, and the central surrogate-model conclusions do not depend on it. The paper's self-citations, e.g. to [12,13,32] for quasi-isodynamic construction and proxies, supply the target definitions and established construction methods; none is invoked as an unverified uniqueness theorem that forces the paper's conclusions. The main weakness is the restricted 8-mode, vacuum, uniform-prior sampling space and the dominance of non-omnigenous equilibria, which limits the generality of the importance ranking to the sampled distribution; that is a generalization and correctness concern, not circularity. Accordingly, no circular step meeting the quoted-reduction standard is present.

Assumptions & free parameters 7 free parameters · 6 assumptions · 0 invented entities

The central conclusions rest on a chain of modeling choices: the VMEC ideal-MHD solver, the f_QS and f_QI proxies, the restriction to 8 Fourier coefficients, the uniform sampling range, and the numerical tolerances. None of these are derived in the paper; they are adopted from the stellarator literature or chosen by hand. The ML models introduce many hyperparameters that are fit to the same database, so the feature-importance results are model-dependent, although the least-important finding is corroborated by a model-free mutual information analysis.

free parameters (7)
  • Fourier coefficient sampling bounds = a=-0.1, b=0.1
    Uniform sampling interval for the eight boundary coefficients chosen by hand; defines the design space explored and thus all feature-importance conclusions.
  • Isolation Forest contamination rate = 0.10
    Chosen to flag outliers; filtering keeps 92.21% of data and affects distribution and correlation analyses.
  • Supervised autoencoder loss weight lambda = 0.5
    Balances reconstruction vs regression loss in Eq. (8); chosen by hand, affects latent space structure.
  • QS threshold for classification label = fQS < 10
    Arbitrary threshold defining the positive class for convergence and omnigenity; affects precision, recall, and the 99.2% accuracy claim.
  • VMEC FTOL and radial grid = FTOL=1e-12, 16 radial points
    Numerical tolerance and grid used for all equilibria; convergence and metric accuracy depend on these choices.
  • ML hyperparameters = Optuna-tuned (e.g., max_depth, num_leaves, learning_rate)
    Hyperparameters of LightGBM, LightGBM LSS, and FFNN are fit to the data via validation; reported performance is conditional on this tuning.
  • Gumbel-Softmax mixture components for LightGBM LSS = M=9 (QS), M=2 (QI), tau=1.0
    Chosen distributional parameters for the probabilistic regression; affect the predicted density shapes and uncertainty estimates.
assumptions (6)
  • domain assumption Ideal MHD equilibrium with nested flux surfaces is computed by VMEC and accurately represents the confining magnetic field.
    Standard tooling assumption; introduced in Section II where VMEC is used to solve J x B = grad p.
  • domain assumption The proxies f_QS and f_QI faithfully quantify omnigenity (alpha-particle confinement).
    Equations (5) and (6) are adopted from Ref. [5] and [32]; the conclusions about influence on omnigenity rely on these proxies.
  • ad hoc to paper mpol=ntor=1 boundary representation with nfp=2 is sufficient to explore omnigenous stellarator design space.
    Section II sets these values; the paper asserts this family can obtain omnigenous configurations but does not compare against known precise designs.
  • ad hoc to paper Uniform random sampling of boundary coefficients produces a representative distribution for training and feature-importance analysis.
    Section III; the density of configurations and correlations are properties of the sampling distribution, not necessarily of the physical design space.
  • ad hoc to paper VMEC convergence with FTOL=1e-12 and 16 radial grid points is sufficient for the metrics used.
    Section II; no convergence study with respect to radial resolution is reported.
  • domain assumption Trained ML models generalize from the sampled database to unseen stellarator shapes.
    The paper's stated purpose is to create fast surrogate models; this assumes the training distribution covers the relevant input space.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Data-Driven Approach to Model the Influence of Magnetic Geometry in the Confinement of Fusion Devices." pith.science (2026). https://pith.science/paper/RXX37W5V

@misc{pith2026250703776,
  author       = {Pith},
  title        = {Pith review of: Data-Driven Approach to Model the Influence of Magnetic Geometry in the Confinement of Fusion Devices},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/RXX37W5V}},
  note         = {Machine review of arXiv:2507.03776}
}
read the original abstract

The design of fusion energy devices involves a balance between competing performance metrics to achieve an energy gain. In stellarators, the geometry is very flexible and involves a large number of free parameters. These can be optimized to achieve good performance. One of the main optimization targets is omnigenity, that is, the confinement of alpha particles stemming from the fusion reactions. In this work, two classes of omnigenous stellarators are studied, namely quasisymmetric and quasi-isodynamic stellarators. The goal is to determine the influence of the geometry on omnigenity, which can lead to greater insight into the design space of stellarators. For this purpose, a database of stellarator configurations is created and analyzed for correlations, pair-wise distributions, and dimensionality reduction using a supervised autoencoder framework. Then, a classification model is trained on this database to predict the convergence of numerical solvers. Finally, two regression models, LightGBM and its probabilistic version, LightGBM LSS, as well as a feed-forward neural network, are trained to predict quasisymmetry and quasi-isodynamiticity and find the design parameters that most influence omnigenity.

Figures

Figures reproduced from arXiv: 2507.03776 by the authors.

Figure 1
Figure 1. FIG. 1. Comparison between a tokamak (left), a quasisymmetric (center), and a quasi-isodynamic stellarator (right). Colors [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. FIG. 2. A set of eight stellarators where a constant value of 0.2 is added to the array [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. FIG. 3. Red: anomalous datapoints as flagged by the Iso [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (11 more)
Figure 4
Figure 4. Figure 4: FIG. 4. Correlation heatmap between physical metrics, both for every point in the database (left) and after filtering (right). [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]
Figure 5
Figure 5. Figure 5: FIG. 5. Top: Marginal distributions of QS, QI, and physics [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]
Figure 6
Figure 6. Figure 6: FIG. 6. Dimensionality reduction of the surface Fourier coefficients using Principal Component Analysis PCA colored by the [PITH_FULL_IMAGE:figures/full_fig_p008_6.png]
Figure 7
Figure 7. Figure 7: FIG. 7. Supervised autoencoder model to reduce the dimensionality of the Fourier coefficients of the boundary and predict [PITH_FULL_IMAGE:figures/full_fig_p009_7.png]
Figure 9
Figure 9. Figure 9: FIG. 9. Feature importance for the classification model to [PITH_FULL_IMAGE:figures/full_fig_p012_9.png]
Figure 10
Figure 10. Figure 10: FIG. 10. Comparison of actual and predicted distributions [PITH_FULL_IMAGE:figures/full_fig_p012_10.png]
Figure 11
Figure 11. Figure 11: FIG. 11. Comparison of actual and predicted distributions [PITH_FULL_IMAGE:figures/full_fig_p013_11.png]
Figure 13
Figure 13. Figure 13: FIG. 13. Comparison of feature importance to quasisymme [PITH_FULL_IMAGE:figures/full_fig_p013_13.png]
Figure 14
Figure 14. Figure 14: FIG. 14. Comparison of feature importance to quasi [PITH_FULL_IMAGE:figures/full_fig_p014_14.png]
Figure 15
Figure 15. Figure 15: FIG. 15. Feature importance of the Fourier coefficients of [PITH_FULL_IMAGE:figures/full_fig_p014_15.png]
Figure 17
Figure 17. Figure 17: FIG. 17. Comparison of actual and predicted distributions [PITH_FULL_IMAGE:figures/full_fig_p015_17.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

55 extracted references · 53 canonical work pages

  1. [1]

    Larger minor radii lead to a larger volume available to yield fusion energy, yielding therefore a large energy gain. The magnetic well [34–36] is a measure of MHD stability, for which V ′′(ψ) < 0, with V (ψ) the volume enclosed by a flux surface of constant ψ, is defined as a 5 Column Name Description RBCm,n and ZBS m,n ∈ [−0.1, 0.1] Fourier coefficients ...

    work page

  2. [2]

    This is expected as it was found that log QI had two peaks in the distri- bution

    The quasi-isodynamic model shows more structure in the data with two clearly defined bands. This is expected as it was found that log QI had two peaks in the distri- bution. A comparison between the true feature values against the reconstructed values for 6 random points in the database is performed in Fig. 7 (bottom). We find that the coefficient ZBS 1,0...

    work page arXiv

  3. [3]

    F. C. Schuller. Disruptions in tokamaks. Plasma Physics and Controlled Fusion, 37(11A), 1995

    work page 1995

  4. [4]

    Y. Xu. A general comparison between tokamak and stellarator plasmas. Matter and Radiation at Extremes , 1(4):192, 2016

    work page 2016

  5. [5]

    Helander

    P. Helander. Theory of plasma confinement in non- axisymmetric magnetic fields. Reports on Progress in Physics, 77(8):087001, 2014

    work page 2014

  6. [6]

    Landreman, B

    M. Landreman, B. Medasani, F. Wechsung, A. Giuliani, R. Jorge, and C. Zhu. SIMSOPT: A flexible framework for stellarator optimization. Journal of Open Source Soft- ware, 6(65):3525, 2021

    work page 2021

  7. [7]

    Landreman and E

    M. Landreman and E. Paul. Magnetic Fields with Pre- cise Quasisymmetry for Plasma Confinement. Physical Review Letters, 128(3):035001, 2022

    work page 2022

  8. [8]

    Jorge, A

    R. Jorge, A. G. Goodman, M. Landreman, J. Rodrigues, and F. Wechsung. Single-stage stellarator optimization: combining coils with fixed boundary equilibria. Plasma Physics and Controlled Fusion , 65(7):074003, 2023

    work page 2023

Show all 55 references
  1. [9]

    Jorge, A

    R. Jorge, A. Giuliani, and J. Loizu. Simplified and flex- ible coils for stellarators using single-stage optimization. Physics of Plasmas , 31(11):112501, 2024

    2024

  2. [10]

    S. A. Henneberg, S. R. Hudson, D. Pfefferl´ e, and P. He- lander. Combined plasma–coil optimization algorithms. Journal of Plasma Physics , 87(2):905870226, 2021

    2021

  3. [11]

    S. P. Hirshman and J. C. Whitson. Steepest-descent moment method for three-dimensional magnetohydrody- namic equilibria. Physics of Fluids , 26(12):3553, 1983

    1983

  4. [12]

    A. H. Boozer. Transport and isomorphic equilibria. Physics of Fluids , 26(2):496, 1983

    1983

  5. [13]

    Nuhrenberg and R

    J. Nuhrenberg and R. Zille. Quasi-helically symmetric toroidal stellarators. Physics Letters A, 129(2):113, 1988

    1988

  6. [14]

    Jorge, G

    R. Jorge, G. G. Plunk, M. Drevlak, M. Landreman, J.-F. Lobsien, K. C. Mata, and P. Helander. A single-field- period quasi-isodynamic stellarator. Journal of Plasma Physics, 88(5):175880504, 10 2022

    2022

  7. [15]

    Goodman, K

    A.G. Goodman, K. Camacho Mata, S.A. Henneberg, R. Jorge, M. Landreman, G.G. Plunk, H.M. Smith, R.J.J. Mackenbach, C.D. Beidler, and P. Helander. Constructing precisely quasi-isodynamic magnetic fields. Journal of Plasma Physics , 89(5):905890504, 2023

    2023

  8. [16]

    G. G. Plunk, M. Landreman, and P. Helander. Di- rect construction of optimized stellarator shapes. Part 3. Omnigenity near the magnetic axis. Journal of Plasma Physics, 85(6):905850602, 2019

    2019

  9. [17]

    D. W. Dudt and E. Kolemen. DESC: A stellarator equi- librium solver. Physics of Plasmas , 27(10):102513, 2020

    2020

  10. [18]

    Curvo, D.R

    P. Curvo, D.R. Ferreira, and R. Jorge. Using deep learn- ing to design high aspect ratio fusion devices. Journal of Plasma Physics , 91(1), 2025

    2025

  11. [19]

    Landreman

    M. Landreman. Mapping the space of quasisymmetric stellarators using optimized near-axis expansion. Journal of Plasma Physics , 88(6):905880616, 2022

    2022

  12. [20]

    Landreman, J

    M. Landreman, J. Y. Choi, C. Alves, P. Balaprakash, R. M. Churchill, R. Conlin, and G. Roberg-Clark. How does ion temperature gradient turbulence depend on magnetic geometry? Insights from data and machine learning. arXiv, 2502.11657(physics.plasma-ph), 2025

    2025 arXiv

  13. [21]

    G. Ke, Q. Meng, T. Finley, T. Wang, W. Chen, W. Ma, Q. Ye, and T. Y. Liu. LightGBM: A highly efficient gradient boosting decision tree. In Advances in Neural Information Processing Systems, volume 2017-December, 2017

    2017

  14. [22]

    Marz and T

    A. Marz and T. Kneib. Distributional Gradient Boosting Machines. arXiv, 2204.00778(stat.ml), 2022

    2022 arXiv

  15. [23]

    S. Murat H. A brief review of feed-forward neural net- works. Communications Faculty Of Science University of Ankara, 50(1), 2006

    2006

  16. [24]

    L. S. Shapley. Cores of convex games. International Jour- nal of Game Theory , 1(1), 1971

    1971

  17. [25]

    Lipovetsky and M

    S. Lipovetsky and M. Conklin. Analysis of regression in game theory approach. Applied Stochastic Models in Business and Industry , 17(4), 2001

    2001

  18. [26]

    S. M. Lundberg and S. I. Lee. A unified approach to interpreting model predictions. In Advances in Neural Information Processing Systems, volume 2017-December, 2017

    2017

  19. [27]

    S. A. Cadena, A. Merlo, E. Laude, A. Bauer, A. Agrawal, M. Pascu, M. Savtchouk, E. Guiraud, L. Bonauer, S. Hudson, and M. Kaiser. ConStellaration: A dataset of QI-like stellarator plasma boundaries and optimization benchmarks. arXiv:2506.19583, arXiv:physics.plasm–ph, 2025

    2025

  20. [28]

    R. Laia. github.com/uwplasma/LightGBM VMEC CONVERGENGE, 2025

    2025

  21. [29]

    R. Laia. Zenodo Dataset https://doi.org/10.5281/zenodo.15008344. Zenodo, 1, 2024

    2024 doi

  22. [30]

    Freidberg

    J. Freidberg. Ideal MHD, volume 9781107006. Cambridge University Press, 2014

    2014

  23. [31]

    S. I. Braginskii. Transport processes in a plasma. Reviews of Plasma Physics , 1:205, 1965

    1965

  24. [32]

    Brizard and T

    A. Brizard and T. Hahm. Foundations of nonlinear gy- rokinetic theory. Reviews of Modern Physics , 79(2):421, 2007

    2007

  25. [33]

    Jorge, W

    R. Jorge, W. Sengupta, and M. Landreman. Construc- tion of quasisymmetric stellarators using a direct coordi- nate approach. Nuclear Fusion, 60(7):076021, 2020

    2020

  26. [34]

    A. G. Goodman, K. Camacho-Mata, Henneberg, S. A., R. Jorge, A. Landreman, G. G. Plunk, H. Smith, R. Mackenbach, and P. Helander. Constructing 17 precisely quasi-isodynamic magnetic fields. arXiv, preprint(physics.plasm):2211.09829, 2022

    2022 arXiv

  27. [35]

    P. A. Figueiredo, R. Jorge, J. Ferreira, and P. Rodrigues. Energetic Particle Tracing in Optimized Quasisymmet- ric Stellarator Equilibria. Journal of Plasma Physics , 90(2):905900207, 2024

    2024

  28. [36]

    C. Mercier. Equilibrium and stability of a toroidal mag- netohydrodynamic system in the neighbourhood of a magnetic axis. Nuclear Fusion, 4(3):213, 1964

    1964

  29. [37]

    Landreman and R

    M. Landreman and R. Jorge. Magnetic well and Mercier stability of stellarators near the magnetic axis. Journal of Plasma Physics , 86(5):905860510, 6 2020

    2020

  30. [38]

    P. Kim, R. Jorge, and W. Dorland. The on-axis mag- netic well and Mercier’s criterion for arbitrary stellarator geometries. Journal of Plasma Physics , 87(2):905870231, 2021

    2021

  31. [39]

    C. R. Harris, K. J. Millman, S. J. van der Walt, R. Gom- mers, P. Virtanen, D. Cournapeau, E. Wieser, J. Tay- lor, S. Berg, N. J. Smith, R. Kern, M. Picus, S. Hoyer, M. H. van Kerkwijk, M. Brett, A. Haldane, J. F. del R ´ ıo, M. Wiebe, P. Peterson, P. G´ erard-Marchant, K. She...

    2020

  32. [40]

    F. T. Liu, K. M. Ting, and Z. H. Zhou. Isolation forest. In Proceedings - IEEE International Conference on Data Mining, ICDM , 2008

    2008

  33. [41]

    G. E. Hinton and R. R. Salakhutdinov. Reducing the dimensionality of data with neural networks. Science, 313(5786), 2006

    2006

  34. [42]

    L. Le, A. Patterson, and M. White. Supervised autoen- coders: Improving generalization performance with un- supervised regularizers. Advances in Neural Information Processing Systems, 31, 2018

    2018

  35. [43]

    X. Ma, J. Sha, D. Wang, Y. Yu, Q. Yang, and X. Niu. Study on a prediction of P2P network loan default based on the machine learning LightGBM and XGboost al- gorithms according to different high dimensional data cleaning. Electronic Commerce Research and Applica- tions, 31, 2018

    2018

  36. [44]

    Natekin and A

    A. Natekin and A. Knoll. Gradient boosting machines, a tutorial. Frontiers in Neurorobotics, 7(DEC), 2013

    2013

  37. [45]

    github.com/microsoft/LightGBM, 2025

    Microsoft. github.com/microsoft/LightGBM, 2025

    2025

  38. [46]

    X. Chai, J. Li, J. Zhao, W. Wang, and X. Zhao. LGB- PHY: An Evaporation Duct Height Prediction Model Based on Physically Constrained LightGBM Algorithm. Remote Sensing, 14(14), 2022

    2022

  39. [47]

    Y. Y. Song and Y. Lu. Decision tree methods: applica- tions for classification and prediction. Shanghai Archives of Psychiatry , 27(2), 2015

    2015

  40. [48]

    E. Jang, S. Gu, and B. Poole. Categorical reparameter- ization with gumbel-softmax. In 5th International Con- ference on Learning Representations, ICLR 2017 - Con- ference Track Proceedings, 2017

    2017

  41. [49]

    Akiba, S

    T. Akiba, S. Sano, T. Yanase, T. Ohta, and M. Koyama. Optuna: A Next-generation Hyperparameter Optimiza- tion Framework. In Proceedings of the ACM SIGKDD International Conference on Knowledge Discovery and Data Mining , 2019

    2019

  42. [50]

    K. V. Rashmi and R. Gilad-Bachrach. DART: Dropouts meet multiple additive regression trees. In Journal of Machine Learning Research, volume 38, 2015

    2015

  43. [51]

    Zhang and Y

    C. Zhang and Y. Ma. Ensemble machine learning: Meth- ods and applications , volume 1. Springer, New York, 2012

    2012

  44. [52]

    Y. Shi, J. Li, and Z. Li. Gradient boosting with piece-wise linear regression trees. In IJCAI International Joint Con- ference on Artificial Intelligence , volume 2019-August, 2019

    2019

  45. [53]

    Q. Meng, G. Ke, T. Wang, W. Chen, Q. Ye, Z. M. Ma, and T. Y. Liu. A communication-efficient parallel algo- rithm for decision tree. In Advances in Neural Informa- tion Processing Systems, 2016

    2016

  46. [54]

    P. A. Estevez, M. Tesmer, C. A. Perez, and J. M. Zurada. Normalized mutual information feature selection. IEEE Transactions on Neural Networks , 20(2), 2009

    2009

  47. [55]

    Svozil, V

    D. Svozil, V. Kvasnicka, and J. Posp ´ ıchal. Introduction to multi-layer feed-forward neural networks. Chemometrics and Intelligent Laboratory Systems , 39(1):43, 1997

    1997

Pith tools

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