Machine-learning surrogate models for nonlinear energetic-particle transport predictions in ITER

Donald A. Spong; Jacobo Varela; Luis Garcia; Yashika Ghai

REVIEW 4 major objections 5 minor 18 references

Machine-learning surrogate models for nonlinear energetic-particle transport predictions in ITER

T0 review · 4 major / 5 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read Trained on nonlinear FAR3d gyrofluid simulations of an ITER steady-state scenario, Gaussian-process and hierarchical neural-network surrogates predict energetic beam and alpha-particle transport fluxes with test $R^2$ around 0.98 while…

desk verdict First ML surrogate for AE-driven EP transport in ITER; solid proof-of-concept with a weak cross-scenario test. read the letter →

arxiv 2608.11058 v1 pith:OBB3SQRJ submitted 2026-08-11 physics.plasm-ph

Yashika Ghai , Donald A. Spong , Jacobo Varela , Luis Garcia This is my paper · ORCID

classification physics.plasm-ph

keywords energetic-particletransportAlfvéneigenmodesmachine-learningsurrogateGaussianprocessneuralnetworkITERFAR3dintegratedmodeling

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 claims that machine-learning surrogates can replace expensive nonlinear simulations of energetic-particle transport in ITER design workflows. The authors train a multitask Gaussian process and a hierarchical neural network on nonlinear FAR3d simulations of Alfvén-eigenmode-driven transport of beam and alpha particles, using a seven-variable local plasma-state representation. Both models reproduce the simulated fluxes on unseen test profiles with coefficients of determination near 0.98 and evaluate all retained transport profiles in seconds instead of 158 wall-clock hours, a speedup of five to six orders of magnitude. If these findings hold, iterative reactor design and scenario optimization could afford repeated transport evaluations that currently require thousands of GPU-hours.

What carries the argument

The load-bearing object is the surrogate mapping $(\Gamma_{\rm beam}, \Gamma_\alpha) = F(\rho, q, \hat{s}, n_{\rm beam}, n_\alpha, \nabla n_{\rm beam}, \nabla n_\alpha)$, which states that an instantaneous local plasma state uniquely determines the transport flux. Two complementary learners realize this mapping: a multitask Gaussian process with a Matérn 3/2 ARD kernel that also receives profile-averaged beam and $\alpha$ densities as global context, and a hierarchical neural network split at $\rho = 0.5$, where a low-$\rho$ network's predicted fluxes are fed as extra inputs to a high-$\rho$ network. Uncertainty is carried by the GP posterior variance and by Monte-Carlo dropout in the NN. A flux-variability analysis, which compares fluxes among neighboring states within a Euclidean distance of 0.3 in the normalized feature space, tests whether the mapping is unique enough for the surrogate formulation to be valid.

What would settle it

Run the trained GP and NN surrogates on new nonlinear FAR3d simulations of an ITER regime not represented in training data, such as a strongly bursting or large-scale relaxation case, and check whether test $R^2$ collapses or median relative $L^2$ error grows well beyond the 7–9% range; alternatively, repeat the flux-variability analysis with time-lagged features (e.g., the flux at the previous dump) and see whether the 10–17% of high-variability states near $\rho \sim 0.35$ become uniquely determined.

Watch

Extended reading notes

Core claim

The central discovery is that the saturated-phase transport flux of energetic particles can, to a good approximation, be treated as a function of the local plasma state $(\rho, q, \hat{s}, n_{\rm beam}, n_\alpha, \nabla n_{\rm beam}, \nabla n_\alpha)$, and that both a multitask Gaussian process and a hierarchical neural network trained on these features from nonlinear FAR3d simulations predict beam and $\alpha$ fluxes on test profiles with $R^2$ values of 0.979 and 0.974 (GP) and 0.982 and 0.975 (NN), respectively. The models reconstruct radial flux profiles with median relative $L^2$ errors of approximately 7–9% on unseen data and reduce the cost of evaluating a full set of transport profiles from 158 wall-clock hours to a few seconds. The authors demonstrate this for an ITER steady-state scenario covering both reversed-shear and monotonic-$q$ profiles, restricting training to the statistically stationary nonlinear saturated phase of the simulations where appreciable transport occurs.

Load-bearing premise

The load-bearing premise is that similar values of the seven plasma-state variables always produce similar transport fluxes, with no dependence on history or on unmeasured quantities; if that fails, the surrogate's accuracy will not generalize beyond its training set.

Editorial extensions

If this is right

Integrated modeling workflows could evaluate AE-driven energetic-particle transport repeatedly in seconds rather than hundreds of wall-clock hours, enabling design scans and uncertainty quantification for ITER and future burning-plasma devices.
The GP surrogate offers consistent global uncertainty estimates and slightly better reconstruction near the transport peak, while the NN surrogate's MC-dropout uncertainty distinguishes early- from late-saturation regimes, so the two models can serve complementary roles.
Because the surrogates learn the transport response directly from nonlinear simulation data, the approach transfers to other simulation codes and other devices as a proof of concept for reduced transport modeling.
The predictive uncertainty estimates provide a confidence measure when the surrogates are queried in plasma regimes sparsely represented in the training data.
The demonstrated accuracy supports extending the methodology to time-dependent transport prediction, with the long-term goal of accelerating nonlinear simulations themselves.

Reading between the lines

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

The authors do not test whether the 10–17% of high-variability states near $\rho \sim 0.35$ become uniquely determined when time-lagged features are added; a direct test would show whether the current seven-variable representation is missing memory effects.
Because the flux-variability analysis flags the strongest transport and gradient regimes as least unique, integrated modeling should treat surrogate outputs there with wider uncertainty margins or resample with additional features.
Disagreement between GP posterior variance and MC-dropout uncertainty could serve as an out-of-distribution detector: when the two uncertainty estimates diverge sharply, a plasma state is likely beyond the training manifold.
The five-to-six-orders-of-magnitude speedup applies only to saturated-phase transport evaluation; using the surrogate to follow linear growth or bursting relaxation would place it outside its training regime.

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.

Desk Editor's Note

read the letter

The thing to know about this paper: it's the first ML surrogate for AE-driven energetic-particle transport in an ITER steady-state scenario, and the basic engineering result holds—two standard methods (multitask GP, hierarchical NN) reproduce FAR3d transport fluxes on a small held-out set with R2 around 0.98 and cut evaluation cost by five to six orders of magnitude. That is a real, useful increment for integrated modeling, not a breakthrough, but not a dud either.

What's genuinely good: the flux-variability analysis is a serious attempt to check whether the local plasma-state parameterization is well-posed, and the authors openly flag that 10–17% of states, concentrated near the transport peak, have local-to-global flux variability above 0.3. They also acknowledge the surrogate is limited to the saturated soft-MHD regime and one ITER scenario. The comparison between GP and NN uncertainty behavior is informative and not overclaimed. The citation pattern looks fine; they build on the FAR3d reference scenario and compare against the relevant reduced models (TGLF-EP, kick model, RBQ, Carlevaro).

The soft spots are the ones your stress-test flagged, and they are real. The test set is four profiles drawn from the same two FAR3d runs used for training. Output dumps are separated by roughly 14,000 timesteps within the same nonlinear trajectories, so the held-out profiles are interpolations along those trajectories, not new equilibria or perturbations. The abstract's claim that the surrogates are 'sufficiently accurate ... to be incorporated into integrated modeling workflows' is stronger than the evidence supports. Also, the GP uses profile-averaged densities as additional global inputs, which further ties predictions to the specific simulation run. The flux-variability threshold of 0.3 is ad hoc, and the 10–17% high-variability states sit exactly in the peak transport region. None of this kills the proof-of-concept, but it means the reported R2 values are not evidence of cross-scenario generalization.

Who benefits: fusion modelers who want a fast transport closure for ITER scenario scans, and anyone building ML surrogates for nonlinear plasma simulations. It deserves a serious referee: the methods are reproducible in principle, the claims are mostly measured, and the limitations are stated. The referee should ask for a test on at least one simulation outside the training runs, ideally a different q-profile or perturbation, and a clearer statement that the current test is interpolation, not extrapolation.

My vote: send it to review, with the expectation that the authors tighten the generalization claims and add a genuinely held-out test if feasible. I'd cite it if I were working on EP transport surrogates.

Referee Report

4 major / 5 minor

Summary. The paper develops Gaussian process and hierarchical neural-network surrogate models that predict beam-ion and alpha-particle transport fluxes from nonlinear FAR3d gyrofluid simulations of an ITER steady-state scenario. The models use a seven-dimensional local plasma-state representation (radius, safety factor, magnetic shear, EP densities, and EP density gradients) to map to the two transport fluxes. A flux-variability analysis is used to argue that this local representation is sufficiently unique over most of the sampled feature space. The surrogates are trained on a dataset of about 9180 spatial samples derived from 20 radial profiles from two FAR3d simulations, with 16 profiles for training and 4 for testing. The reported test-set R² values are 0.9791/0.9737 (GP) and 0.9823/0.9752 (NN) for beam/alpha fluxes, and the evaluation time is reduced by five to six orders of magnitude relative to a full nonlinear simulation. The paper frames the result as a proof of concept for replacing repeated transport evaluations in integrated modeling workflows.

Significance. If the central claims hold, this would be a useful and novel proof of concept: the first ML surrogate for AE-driven energetic-particle transport, with explicit uncertainty quantification and a substantial speedup. The comparative study of GP and hierarchical NN uncertainty behavior is also informative. However, the significance is tempered by the narrow data basis: the surrogates are trained and tested on two simulations of one ITER steady-state scenario, and the test set consists of only four profiles drawn from those same two runs. The claimed predictive accuracy is therefore an interpolation result within the training manifold rather than demonstrated generalization to new equilibria, perturbations, or operating scenarios. The paper is appropriately cautious in its conclusions, but the abstract's statement that the surrogate is 'sufficiently accurate and computationally efficient to be incorporated into future integrated modeling workflows' goes beyond what the current evidence strictly supports.

major comments (4)

[§2 and §5.1 (Table 3)] The test set consists of four profiles taken from the same two FAR3d runs used for training, with output dumps separated by about 14,000 timesteps within the nonlinear saturated phase. Consequently, the test profiles lie on the same temporal trajectories as the training profiles, and the reported test R² values of 0.979–0.982 demonstrate interpolation along known simulation trajectories rather than predictive skill for a new plasma state. For the GP, the additional profile-averaged densities ⟨n_beam⟩ and ⟨n_alpha⟩ are dump-level global inputs, which further tie test predictions to the specific source run. The paper should either add a hold-out scenario evaluation (e.g., train on one q-profile and test on the other, or leave out a full simulation) or explicitly restrict the central claim to interpolation within the sampled manifold. As written, the evidence does not support the abstract's implication that the surrogates are ready for deployment in integrated modeling of unseen ITER states.
[§3 (Fig. 3 and surrounding text)] The flux-variability analysis is load-bearing for the uniqueness of the mapping in Eq. (3), but it leaves a non-negligible fraction of the data unexplained: approximately 10% of beam states and 17% of alpha states have local-to-global flux variability above the 0.3 threshold, concentrated near the peak transport region (ρ ≈ 0.35) and at intermediate-to-high gradient strengths. These are precisely the states that contribute most to transport, and the neighborhood radius of 0.3 is selected based on mean nearest-neighbor distances rather than derived from an objective criterion. The paper should quantify how the surrogate prediction errors on these high-variability states compare with errors on the rest of the dataset, and should discuss whether additional features (e.g., mode amplitudes or history-dependent variables) would be needed to make the mapping unique on this subset. Without this, the statement that the representation 'provides a sufficient unique parameterization over most of the sampled feature space' is too strong.
[§5.2 and §5.3] The claim that 'neither model exhibits significant overfitting' is based on the similarity between training and testing L2 errors (Table 3). Because the test profiles come from the same two simulations as the training profiles, this comparison does not actually rule out overfitting to the simulation-specific trajectory. A more convincing check would be leave-one-simulation-out cross-validation or an evaluation on a simulation with a different initial condition or perturbation. The discussion of outlier profiles in §5.3 is useful, but it does not substitute for an independent test set.
[§4.1] The hierarchical NN architecture is not fully specified. The high-ρ network uses 'the mean beam and alpha-particle transport fluxes predicted by the low-ρ network' as additional inputs, but it is unclear over which radial range or set of low-ρ points this mean is taken, and how the low-ρ network is applied when the high-ρ network is evaluated at inference time. This matters because the high-ρ network's inputs depend on the low-ρ network's outputs, creating a sequential dependency not described in enough detail for reproducibility. The manuscript should provide an explicit computational recipe for the hierarchical evaluation.

minor comments (5)

[§2] The text alternates between 'normal-shear' and 'monotonic q-profile' when referring to the second simulation. Please use a single consistent term throughout.
[§5.2, Eq. (4)] The notation Γpred(ρ) and Γtrue(ρ) is introduced in Eq. (4) but the integration variable and the normalization are not fully defined; specify that ρ is the square-root of normalized toroidal flux and that the integral is over the radial domain.
[§5.3] The 'distance correlation coefficient' is used to characterize the flux-gradient relationship, but no definition or reference is given. Please add a brief definition or citation.
[Figure 3(b)] The figure caption says the 'average number of similar cross-profile states' is shown, but the vertical axis label and the text would be clearer if the units (count per radial bin) were made explicit.
[§2, 'Training vs testing data split'] The sentence 'For the normal-shear simulations, we retained 9 output dumps' appears to refer to the monotonic q-profile case; please reconcile the terminology as noted above.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the surrogate is an empirical fit to FAR3d outputs with held-out test profiles, and no load-bearing step reduces to its own inputs or to a self-citation chain.

full rationale

The paper's derivation chain is an empirical surrogate construction, not a first-principles derivation. The GP and NN are trained on radial profiles from two FAR3d nonlinear simulations and evaluated on four held-out profiles (two per q-profile) that were excluded by explicit profile-wise partitioning (Section 2). The test R^2 values in Table 3 therefore measure interpolation over radius and time within the same two simulated scenarios, but they do not reduce to the fitted values by construction, because the test profiles were not used to set model weights or hyperparameters. The local-state ansatz in Eq. (3) is checked, rather than assumed, by the flux-variability analysis in Section 3, which honestly reports that 10% (beam) and 17% (alpha) of states have local-to-global variability above 0.3; this is a validity limitation in strong-transport regions, not a circular step. FAR3d is cited to Refs. [3,7-10], including prior work by the authors, but these citations supply the simulation code, dataset, and JET benchmark, not an imported uniqueness theorem or an unexamined ansatz. The prediction of FAR3d transport is a standard supervised-learning approximation of a fixed simulation mapping, with no load-bearing reduction of the result to its own inputs. Remaining concerns about generalization to new equilibria or scenarios are a scope limitation, not circularity.

Assumptions & free parameters 5 free parameters · 4 assumptions · 0 invented entities

The surrogate models are empirical fits to FAR3d simulation outputs. The central claim's validity depends on the fidelity of the FAR3d code, the completeness of the chosen feature representation, and the representativeness of the small test set. Key free parameters include the feature set, the neighborhood radius (0.3) used in the flux-variability analysis, the NN architecture and hyperparameters, and the GP kernel hyperparameters fitted to the training data.

free parameters (5)

Feature-space neighborhood radius (Euclidean distance) for flux-variability analysis = 0.3
Chosen based on mean nearest-neighbor distance (0.16-0.23) in normalized feature space; ad hoc threshold that determines the fraction of states classified as high-variability.
Hierarchical NN radial boundary = rho = 0.5
Chosen to split inner and outer plasma regions; no optimization or sensitivity analysis reported for this boundary.
NN architecture and hyperparameters (learning rate, epochs, hidden width, dropout rate, loss weighting) = lr 5e-4, 3000 epochs, 64 neurons per layer, dropout 10%, weighted MSE
Hand-chosen values; the weighting scheme for the MSE loss is not specified precisely, so the effective loss is not fully reproducible.
GP kernel hyperparameters (ARD length scales, multitask covariance, observation noise) = optimized by marginal likelihood maximization
Fitted to training data; these are free parameters of the surrogate model.
Flux normalization factor = 1e-5 times characteristic flux scale
Introduced for numerical conditioning; changes only target magnitude, not relative variation.

assumptions (4)

domain assumption FAR3d gyrofluid simulations faithfully represent AE-driven energetic-particle transport in ITER
The surrogate inherits all simulation errors; the paper cites JET validation (Refs [7-9]) but no burning-plasma validation exists.
domain assumption Instantaneous transport flux is a unique function of the local plasma state (Eq. 3), with no significant memory or hysteresis
Stated in Section 2 ('the surrogate model assumes that the instantaneous transport fluxes can be parameterized by the current plasma state'); tested in Section 3 but not proven.
domain assumption The nonlinear saturated (soft-MHD) regime is the relevant regime for integrated modeling
The training data are drawn only from this regime; bursting or large-scale relaxation regimes are explicitly excluded (Sections 2 and 6).
domain assumption The held-out test profiles from the same two simulations are representative of the broader ITER operating space
Only 4 test profiles from the same reversed-shear and monotonic-q simulations are used; the paper limits claims to the training database but the abstract implies broader applicability.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Machine-learning surrogate models for nonlinear energetic-particle transport predictions in ITER." pith.science (2026). https://pith.science/paper/OBB3SQRJ

@misc{pith2026260811058,
  author       = {Pith},
  title        = {Pith review of: Machine-learning surrogate models for nonlinear energetic-particle transport predictions in ITER},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/OBB3SQRJ}},
  note         = {Machine review of arXiv:2608.11058}
}

read the original abstract

Fast and accurate prediction of energetic-particle transport driven by Alfv\'en eigenmode (AE) instabilities is essential for integrated modeling workflows used in the design and optimization of burning plasma fusion reactors. In this work, we develop machine-learning-based surrogate models for rapid prediction of energetic beam and alpha-particle transport fluxes, together with predictive uncertainty estimates, for an ITER steady-state scenario. Two complementary surrogate methodologies, Gaussian process (GP) regression and hierarchical neural networks (NNs), are trained using nonlinear FAR3d gyrofluid simulations of energetic-particle transport. A flux-variability analysis demonstrates that the selected plasma-state representation provides a sufficiently unique parameterization of the nonlinear transport response over most of the sampled feature space, thereby justifying the surrogate formulation. Both surrogate models reproduce the nonlinear transport fluxes with high predictive accuracy while reducing the computational cost of transport evaluation by approximately five to six orders of magnitude relative to direct nonlinear FAR3d simulations. Although the two approaches achieve comparable predictive accuracy, they exhibit distinct uncertainty characteristics: the GP provides more consistent global uncertainty estimates, whereas the NN more clearly distinguishes between different transport regimes. This work establishes a proof of concept for developing machine-learning surrogate models of energetic-particle transport that are sufficiently accurate and computationally efficient to be incorporated into future integrated modeling workflows.

Figures

Figures reproduced from arXiv: 2608.11058 by the authors.

**Figure 2.** Variation of alpha particle transport flux (Γ [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗

**Figure 3.** Assessment of the surrogate state representation using nearest-neighbor anal [PITH_FULL_IMAGE:figures/full_fig_p009_3.png] view at source ↗

Figures from the paper (6 more)

**Figure 4.** Figure 4: Scatter plots showing energetic alphas and beam particle fluxes predicted by [PITH_FULL_IMAGE:figures/full_fig_p012_4.png]

**Figure 5.** Figure 5: Scatter plots showing energetic alphas and beam particle fluxes predicted [PITH_FULL_IMAGE:figures/full_fig_p013_5.png]

**Figure 6.** Figure 6: Comparison of the residual transport flux errors, defined as Γ [PITH_FULL_IMAGE:figures/full_fig_p014_6.png]

**Figure 7.** Figure 7: Boxplots of the relative L 2 error, computed using Eq. 4, for beam and alphaparticle flux predictions obtained from the GP (top row) and NN (bottom row) surrogate models. The colored box spans the interquartile range (IQR), corresponding to the 25th–75th percentiles …

**Figure 8.** Figure 8: Plots of mean predictive uncertainty in flux profile calculation for ITER steady [PITH_FULL_IMAGE:figures/full_fig_p018_8.png]

**Figure 9.** Figure 9: Representative alpha-particle transport flux profiles predicted by the GP (top [PITH_FULL_IMAGE:figures/full_fig_p019_9.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

18 extracted references · 17 canonical work pages

[1]

Mayoral, R

M.-L. Mayoral, R. Buttery, T. T. C. Jones, V. Kiptily, S. Sharapov, M. J. Mantsi- nen, S. Coda, O. Sauter, L.-G. Eriksson, F. Nguyen, D. N. Borba, A. M¨ uck, S. D. Pinches, J.-M. Noterdaeme, and JET-EFDA Contributors. Studies of burning plasma physics in the joint european torus.Physics of Plasmas, 11(5):2607–2615, 05 2004

work page 2004
[2]

Salewski, D.A

M. Salewski, D.A. Spong, P. Aleynikov, R. Bilato, B.N. Breizman, S. Briguglio, H. Cai, L. Chen, W. Chen, V.N. Duarte, R.J. Dumont, M.V. Falessi, M. Fitzger- ald, E.D. Fredrickson, M. Garc ´ ıa-Mu˜ noz, N.N. Gorelenkov, T. Hayward-Schneider, W.W. Heidbrink, M.J. Hole, Ye.O. Kazakov, V.G. Kiptily, A. K¨ onies, T. Kurki- Suonio, Ph. Lauber, S.A. Lazerson, Z....

work page 2025
[3]

Varela, D

J. Varela, D. Spong, L. Garcia, Y. Ghai, J. Ortiz, and F AR3d project collaborators . Stability optimization of energetic particle driven modes in nuclear fusion devices: the far3d gyro-fluid code.Frontiers in Physics, Volume 12 - 2024, 2024

work page 2024
[4]

Liu, H.R

Z.Y. Liu, H.R. Qiu, G.Y. Fu, Y. Xiao, Y.C. Chen, Z.J. Wang, and Y.X. Wei. Pre- diction of fishbone linear instability in tokamaks with machine learning methods. Nuclear Fusion, 65(1):016007, nov 2024

work page 2024
[5]

G. Dong, X. Wei, J. Bao, G. Brochard, Z. Lin, and W. Tang. Deep learning based surrogate models for first-principles global simulations of fusion plasmas.Nuclear Fusion, 61(12):126061, nov 2021

work page 2021
[6]

A. M. Bruncrona, A. Kit, A. E. J¨ arvinen, S. Saarelma, L. Frassinetti, and JET Con- tributors. Machine learning surrogate model for ideal peeling–ballooning pedestal mhd stability.Physics of Plasmas, 32(9):092501, 09 2025

work page 2025
[7]

Solano, ˇZiga ˇStancar, Jacobo Varela, Matteo Baruzzo, Emily Belli, Phillip J

Jeronimo Garcia, Yevgen Kazakov, Rui Coelho, Mykola Dreval, Elena de la Luna, Emilia R. Solano, ˇZiga ˇStancar, Jacobo Varela, Matteo Baruzzo, Emily Belli, Phillip J. Bonofiglo, Jeff Candy, Costanza F. Maggi, Joelle Mailloux, Samuele Mazzi, Jef Ongena, Juan R. Ruiz, Michal Poradzinski, Sergei Sharapov, David Zarzoso, and JET contributors. Stable deuterium...

work page 2024
[8]

Varela, J

J. Varela, J. Garcia, S. Mazzi, Y. Kazakov, Z. Stancar, M. Baruzzo, J. Ongena, D.A. Spong, L. Garcia, Y. Ghai, D. Zarzoso, J. Ortiz, M. Poradzinski, S. Shara- pov, M. Fitzgerald, B. Breizman, F. Waelbroeck, S. Menmuir, H.J. Sun, D. Kos, A. Boboc, N.C. Hawkes, JET Contributors, and the EUROfusion Tokamak Ex- ploitation team. Analysis of the linear and nonl...

work page 2025

Show all 18 references

[9]

Varela, J

J. Varela, J. Garcia, D. Zarzoso, S. Mazzi, Y. Kazakov, Z. Stancar, M. Baruzzo, J. Ongena, D.A. Spong, L. Garcia, Y. Ghai, J. Ortiz, M. Poradzinski, S. Sharapov, M. Fitzgerald, B. Breizman, D. del Castillo-NEgrete, F. Waelbroeck, S. Menmuir, H.J. Sun, D. Kos, A. Boboc, N.C. Ha...

2026
[10]

Spong, Y

D.A. Spong, Y. Ghai, J. Varela, and L. Garcia. Nonlinear alfv´ en instability simula- tion and ep transport for iter reversed shear (steady-state) and monotonic q-profile regimes*.Nuclear Fusion, 65(11):112004, oct 2025. 21

2025
[11]

Zweben, R.V

S.J. Zweben, R.V. Budny, D.S. Darrow, S.S. Medley, R. Nazikian, B.C. Stratton, E.J. Synakowski, and G. Taylor for the TFTR Group. Alpha particle physics experiments in the tokamak fusion test reactor.Nuclear Fusion, 40(1):91, jan 2000

2000
[12]

He Sheng, R. E. Waltz, and G. M. Staebler. Alfv´ en eigenmode stability and critical gradient energetic particle transport using the trapped-gyro-landau-fluid model.Physics of Plasmas, 24(7):072305, 06 2017

2017
[13]

Bass and R.E

E.M. Bass and R.E. Waltz. Prediction of alfv´ en eigenmode energetic particle trans- port in iter scenarios with a critical gradient model.Nuclear Fusion, 60(1):016032, dec 2019

2019
[14]

Podest` a, L

M. Podest` a, L. Bard´ oczi, C.S. Collins, N.N. Gorelenkov, W.W. Heidbrink, V.N. Duarte, G.J. Kramer, E.D. Fredrickson, M. Gorelenkova, D. Kim, D. Liu, F.M. Poli, M.A. Van Zeeland, and R.B. White. Reduced energetic particle transport models enable comprehensive time-dependent ...

2019
[15]

Gorelenkov, V.N

N.N. Gorelenkov, V.N. Duarte, M. Podesta, and H.L. Berk. Resonance broad- ened quasi-linear (rbq) model for fast ion distribution relaxation due to alfv´ enic eigenmodes.Nuclear Fusion, 58(8):082016, jun 2018

2018
[16]

Lauber, Z Lu, and X Wang

N Carlevaro, G Meng, G Montani, F Zonca, T Hayward-Schneider, Ph. Lauber, Z Lu, and X Wang. One dimensional reduced model for iter relevant energetic particle transport.Plasma Physics and Controlled Fusion, 64(3):035010, jan 2022

2022
[17]

Carl Edward Rasmussen and Christopher K. I. Williams.Gaussian Processes for Machine Learning. MIT Press, 2006

2006
[18]

Dropout as a bayesian approximation: Rep- resenting model uncertainty in deep learning

Yarin Gal and Zoubin Ghahramani. Dropout as a bayesian approximation: Rep- resenting model uncertainty in deep learning. InInternational Conference on Ma- chine Learning, 2016. 22

2016

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.

[1] [1]

Mayoral, R

M.-L. Mayoral, R. Buttery, T. T. C. Jones, V. Kiptily, S. Sharapov, M. J. Mantsi- nen, S. Coda, O. Sauter, L.-G. Eriksson, F. Nguyen, D. N. Borba, A. M¨ uck, S. D. Pinches, J.-M. Noterdaeme, and JET-EFDA Contributors. Studies of burning plasma physics in the joint european torus.Physics of Plasmas, 11(5):2607–2615, 05 2004

work page 2004

[2] [2]

Salewski, D.A

M. Salewski, D.A. Spong, P. Aleynikov, R. Bilato, B.N. Breizman, S. Briguglio, H. Cai, L. Chen, W. Chen, V.N. Duarte, R.J. Dumont, M.V. Falessi, M. Fitzger- ald, E.D. Fredrickson, M. Garc ´ ıa-Mu˜ noz, N.N. Gorelenkov, T. Hayward-Schneider, W.W. Heidbrink, M.J. Hole, Ye.O. Kazakov, V.G. Kiptily, A. K¨ onies, T. Kurki- Suonio, Ph. Lauber, S.A. Lazerson, Z....

work page 2025

[3] [3]

Varela, D

J. Varela, D. Spong, L. Garcia, Y. Ghai, J. Ortiz, and F AR3d project collaborators . Stability optimization of energetic particle driven modes in nuclear fusion devices: the far3d gyro-fluid code.Frontiers in Physics, Volume 12 - 2024, 2024

work page 2024

[4] [4]

Liu, H.R

Z.Y. Liu, H.R. Qiu, G.Y. Fu, Y. Xiao, Y.C. Chen, Z.J. Wang, and Y.X. Wei. Pre- diction of fishbone linear instability in tokamaks with machine learning methods. Nuclear Fusion, 65(1):016007, nov 2024

work page 2024

[5] [5]

G. Dong, X. Wei, J. Bao, G. Brochard, Z. Lin, and W. Tang. Deep learning based surrogate models for first-principles global simulations of fusion plasmas.Nuclear Fusion, 61(12):126061, nov 2021

work page 2021

[6] [6]

A. M. Bruncrona, A. Kit, A. E. J¨ arvinen, S. Saarelma, L. Frassinetti, and JET Con- tributors. Machine learning surrogate model for ideal peeling–ballooning pedestal mhd stability.Physics of Plasmas, 32(9):092501, 09 2025

work page 2025

[7] [7]

Solano, ˇZiga ˇStancar, Jacobo Varela, Matteo Baruzzo, Emily Belli, Phillip J

Jeronimo Garcia, Yevgen Kazakov, Rui Coelho, Mykola Dreval, Elena de la Luna, Emilia R. Solano, ˇZiga ˇStancar, Jacobo Varela, Matteo Baruzzo, Emily Belli, Phillip J. Bonofiglo, Jeff Candy, Costanza F. Maggi, Joelle Mailloux, Samuele Mazzi, Jef Ongena, Juan R. Ruiz, Michal Poradzinski, Sergei Sharapov, David Zarzoso, and JET contributors. Stable deuterium...

work page 2024

[8] [8]

Varela, J

J. Varela, J. Garcia, S. Mazzi, Y. Kazakov, Z. Stancar, M. Baruzzo, J. Ongena, D.A. Spong, L. Garcia, Y. Ghai, D. Zarzoso, J. Ortiz, M. Poradzinski, S. Shara- pov, M. Fitzgerald, B. Breizman, F. Waelbroeck, S. Menmuir, H.J. Sun, D. Kos, A. Boboc, N.C. Hawkes, JET Contributors, and the EUROfusion Tokamak Ex- ploitation team. Analysis of the linear and nonl...

work page 2025