Pith. sign in

REVIEW 4 major objections 7 minor 40 references

Physics-Gated Visual Prediction of MARFE on the HL-3 Tokamak

T0 review · 4 major / 7 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read A physics-gated Neural ODE can warn of MARFE worsening 40 ms ahead on HL-3, with reported AUC 0.969 and a 1 ms runtime.

desk verdict A real deployed MARFE warning system on HL-3, but the headline numbers measure anticipation of the model's own cleaned video features, not independently verified MARFE onset. read the letter →

arxiv 2510.24347 v2 pith:6DJC23NC submitted 2025-10-28 physics.plasm-ph

Qianyun Dong , Rongpeng Li , Zongyu Yang , Niannian Wu , Fan Xia , Liang Liu , Zhifeng Zhao , Wulyu Zhong This is my paper · ORCID
classification physics.plasm-ph PACS 52.55.Fa
keywords MARFEdisruptionpredictiontokamakHL-3NeuralODEphysics-informedmachinelearningvisible-lightimagingexpectation-maximization
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 sets out to show that a tokamak can be given a reliable, real-time warning that a MARFE—the edge radiation instability that often precedes density-limit disruptions—is about to worsen, using only a visible-light camera and routinely available plasma signals. Its pipeline first cleans noisy camera-derived MARFE features with a weighted expectation-maximization step whose prior is a physics score built from $n_e$, $T_e$, $f_G$, and time, then feeds the cleaned features into a continuous-time Neural ODE whose dynamics are gated by $f_G$ and $T_e$ and integrated 40 ms forward. On a shot-level held-out set the paper reports an Area Under the Curve (AUC) of 0.969 in the experimental section (0.981 in the abstract) and an $F_1$ score of 0.862, and the optimized deployed inference service runs within the 1 ms control-cycle budget. If these claims hold, the system offers a lead time long enough for shape-control setpoints to be computed before a MARFE worsens, which is what proactive mitigation requires.

What carries the argument

The load-bearing machinery is the pair formed by the physics-scored EM label refiner and the physics-gated Neural ODE, where the ODE is a neural network that parameterizes the derivative of a latent state. In the refiner, a physics consistency score $s_i$ is computed from indicator thresholds on $n_e$, $T_e$, $f_G$, and discharge time, and a weighted EM algorithm fits a two-component Gaussian mixture with $s_i$ as a Bayesian prior; the converged responsibilities $\gamma(z_i,\mathrm{pos})$ become the refined label $\hat{y}_i$, which is then used to zero out inconsistent visual features. The predictor evolves a latent state $h(t)$ by $dh/dt=f_\theta(h)+g(f_G,T_e)\,f_\phi(h)$, with $g$ a sigmoid gate activating as $f_G$ exceeds $f_G^{\mathrm{thr}}$ and $T_e$ falls below $T_e^{\mathrm{thr}}$; integrating from $t_0$ to $t_0+40\,\mathrm{ms}$ yields the future latent state decoded into three zone-wise worsening probabilities. The 40 ms horizon is matched to the actuator-response budget, and the deployment fits the 1 ms control cycle.

What would settle it

Take the trained model and its EM label pipeline to a set of shots in which MARFE onset is marked independently—by high-field-edge peaking in bolometric radiated power, by MARFE-zone spectroscopic line emission, or by expert review of raw video—and require that prediction threshold crossings lead those markers by at least 40 ms at the claimed false-positive rate. If the aligned lead time does not reproduce, the AUC mostly certifies label self-consistency, not prediction of MARFE.

Watch

Extended reading notes

Core claim

The central claim is that MARFE worsening on HL-3 is predictable 40 ms in advance by evolving a learned latent state in continuous time, with extra dynamics switched on by a sigmoid physics gate $g(f_G,T_e)=\sigma\big(k_n(f_G-f_G^{\mathrm{thr}})-k_T(T_e-T_e^{\mathrm{thr}})\big)$ as the plasma approaches the MARFE-prone region. The training labels are not raw camera thresholds: an EM algorithm over a two-component Gaussian mixture, using each sample's physics score $s_i$ as a sample-specific prior, refines the noisy binary labels and then zeros out the visual features the EM judges inconsistent. The full model achieves AUC 0.969 and $F_1=0.862$; ablations show training without refined labels drops AUC to 0.908, replacing the ODE with a Bi-LSTM drops it to 0.935, and removing the gate leaves global AUC almost unchanged but reduces the true-positive rate at low false-positive rates. The paper further claims that the deployed service completes the whole pipeline in about 1 ms, matching the HL-3 control cycle, and outputs six shape-target quantities for the controller.

Load-bearing premise

The load-bearing premise is that the EM-refined 'worsening' labels really correspond to MARFE onset; since those labels are thresholded future jumps in the same cleaned camera features the model is trained on, and the cleaning is driven by the same physics measurements that gate the model, the reported lead time and accuracy could reflect self-consistency rather than predictive skill unless the labels are independently verified.

Editorial extensions

If this is right

  • A 40 ms-ahead warning is compatible with the HL-3 control cycle: the shape-target calculation completes before the worsening interval begins, so the warning is actionable rather than retrospective.
  • In the low-false-alarm region that controllers demand, the physics gate provides a higher true-positive rate than the ungated model at the same false-positive rate, meaning fewer spurious interventions.
  • Training on EM-refined labels rather than raw camera thresholds raises the reported AUC from 0.908 to 0.969, so the label-refinement step, not the ODE alone, is responsible for much of the accuracy.
  • Because the core is a continuous-time ODE, the same trained model can be queried at horizons other than 40 ms without retraining, which would allow the warning horizon to be tuned to a device's actuator response.
  • The deployed pipeline logs raw indicators and probabilities into the HL-3 data systems, so the predictor doubles as a diagnostic archive for post-shot MARFE studies.

Reading between the lines

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

  • Because the worsening label is defined as a thresholded future jump in the same cleaned visual features the model consumes, the reported lead time measures internal label self-consistency unless the refined labels are independently confirmed; a bolometer- or spectroscopy-based MARFE-onset set would settle that.
  • The gate thresholds and score weights are fitted to HL-3 training data, so porting the architecture to another tokamak would require re-estimating them; the claim of improved generalization is therefore plausible but device-specific.
  • The paper stops at open-loop deployment; the practical payoff would be demonstrated by closed-loop tests in which the six shape targets are fed to the plasma control system and MARFE onset is suppressed or delayed relative to the open-loop baseline.
  • Because $f_G$ is computed from core rather than line-averaged density, its values sit above the usual density-limit-normalized literature values; cross-device comparisons should use a consistent density definition before treating a given $f_G$ threshold as transferable.
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

4 major / 7 minor

Summary. The paper presents a real-time MARFE prediction system for the HL-3 tokamak. The authors propose a three-stage label refinement pipeline that converts noisy CCD-camera area features into cleaned features using a physics-scored EM algorithm, and then train a physics-gated Neural ODE to predict a 40 ms-ahead 'MARFE worsening' label per poloidal zone. The model is evaluated on a held-out shot split with reported AUC 0.969 or 0.981 (abstract and body disagree), and the deployed TensorRT service is claimed to run within a 1 ms control-cycle budget. The central claim is that the system provides an actionable 40 ms warning of MARFE worsening, with a median lead time of +36 ms reported in the abstract.

Significance. The paper addresses a practically important problem—real-time MARFE warning for tokamak control—and the engineering contributions are real: a deployed TensorRT service with sub-millisecond inference, a physics-gated Neural ODE that naturally treats the 40 ms horizon as continuous-time evolution, and a large HL-3 dataset. The EM-based label refinement is a thoughtful attempt to suppress visual artifacts, and the paper honestly acknowledges in §3.1 the absence of a comprehensive manually-labeled ground truth. If the label circularity is resolved—for example by independent validation against bolometry or expert labels—the approach would be a useful addition to the disruption-avoidance toolbox. However, as it stands, the headline AUC and lead-time numbers cannot be distinguished from self-consistency of the labeling pipeline, and the abstract/body inconsistencies undermine confidence in the exact performance claims.

major comments (4)
  1. [§2.3, §2.4.2, §3.1] The central claim of 40 ms-ahead MARFE prediction is not supported by an independent ground truth. The worsening label b_j(t) in Eq. (8) is defined as a thresholded future jump of the cleaned features m'_j, which are exactly the model inputs after EM-based zeroing in Eq. (7). Those cleaned features are zeroed whenever the EM refinement (Eqs. (2)–(6)), driven by the same plasma variables n_e, T_e, f_G used in the physics gate g(f_G,T_e) of Eq. (10), decides that no MARFE is present. Consequently, the reported AUC/F1 measure the model's ability to anticipate a jump in its own cleaned input, and the abstract's claim of a '+36 ms median label-aligned lead time' is a statement about self-consistency of the labeling pipeline rather than verified predictive skill. The paper explicitly acknowledges in §3.1 the absence of a comprehensive manually-labeled ground truth and offers only qualitative checks (correction times, Shot #11522). To establish the headline claim, the authors should either (i) validate the refined labels against an independent diagnostic (e.g., bolometer/radiated-power signals or expert-labeled frames for a subset of shots), or (ii) reframe the claims as predicting the refined pseudo-label and add a control baseline that directly thresholds the slope or future jump of m'_j to show the Neural ODE adds skill beyond trivial extrapolation.
  2. [Abstract vs. §3.2.1, Table 3, Figure 8] The numerical reporting is internally inconsistent. The abstract reports AUC=0.981, F1=0.840 on a 'frozen 140-shot held-out test set', while §3.2.1 and Figure 8 report AUC=0.969 (and the text also mentions 0.959), Table 3 reports AUC=0.9685 and F1=0.8618, and §3 describes an 8:1:1 split of 701 shots, which implies roughly 70 test shots rather than 140. The ablation text gives 'without refined labels' AUC=0.868 in one place and 0.908 in the figure caption. These discrepancies make it impossible to identify the official evaluation protocol or the size of the test set, and they must be resolved before the reported performance can be considered reproducible.
  3. [§2.4.2] Even if the label circularity were accepted as a deliberate soft-target design, the paper does not compare against a trivial predictor that uses the same 40 ms future jump directly. Since b_j(t) is, up to fitted thresholds and a slope condition, essentially an indicator that m'_j will increase by more than θ_j over the next 40 ms, a baseline that thresholds the recent slope or the current value of m'_j could achieve high AUC without any learned dynamics. Reporting such a baseline would clarify whether the Neural ODE's continuous-time modeling (and the physics gate) adds genuine predictive information beyond the label construction.
  4. [§3.2.1] The abstract's claim of a 'median label-aligned lead time of +36 ms' is not supported by any systematic analysis in the body. The only lead-time evidence is a single illustrative shot (#6255) with '>20 ms' lead time. If the median lead time is computed over the test set, the corresponding distribution (or at least the median and an interval) should be reported in the main text; otherwise the abstract's number is unverifiable.
minor comments (7)
  1. [§2.4.4, Eq. (11)] The uncertainty weights are written as σ_j in the equation but σ_i in the surrounding text; please unify the notation.
  2. [Table 5 vs. §2.3] In Table 5, the EM feature vector is listed as [n_e, T_e, q_95, t]_i, while Eqs. (2)–(3) and the text define x_i = [n_e, T_e, f_G, t]_i. Please correct the inconsistency.
  3. [Figure 6] The figure caption appears after a corrupted Unicode dump in the manuscript; the figure rendering must be fixed so the reader can actually see the time distribution of corrections.
  4. [§2.1] The observed ranges for n_e and T_e are given as [−3, 15] and [−1, 13]; negative values are physically suspicious and the units or offset should be clarified or corrected.
  5. [§3.2.1] Please provide the hyperparameters of the sequence encoder, f_θ, f_ϕ, the ODE solver settings, and the training schedule; without these, the ablation and deployment results cannot be reproduced by the community.
  6. [Throughout] The terms 'physics-gated' and 'physics-constrained' are used interchangeably; pick one term and use it consistently.
  7. [§1] The sentence 'This work lays a very foundation' contains a typo and should read 'lays a foundation'.

Circularity Check

2 steps flagged · score 6.0 of 10

Reported AUC/F1 and lead time measure self-consistency: the worsening label in Eq. (8) is a thresholded future jump of the cleaned features that are themselves the model inputs, with thresholds fit to training data; no independent MARFE ground truth is used.

  1. fitted input called prediction [Section 2.4.2, Eq. (8), and Table 2 (model inputs)]
    "a notable jump ∆m′_j(t)= m′_j(t+40ms)−m′_j(t) in the future 40 ms ... b_j(t)=1 if [ (∆m^inst_j(t)>θ^inst_j)∧(G_j(t)>0) or (∆m^inst_j(t)>c·θ^inst_j) ], ... θ_i is a pre-defined threshold for a “notable jump” specific to each region i, determined empirically from the training data."

    The cleaned features (m′_U,m′_M,m′_L) are listed as model inputs in Table 2. Eq. (8) defines the ground-truth label b_j(t) as a thresholded future increment of exactly those same channels, with θ_j fit to training data. Thus the reported AUC/F1/lead time measure whether the model can anticipate a deterministic temporal transform of its own input, not whether it detects an independently verified MARFE onset. The paper itself notes in §3.1 that no comprehensive manually-labeled ground truth exists. This makes the headline prediction target self-definitional.

  2. self definitional [Section 2.2, Eqs. (1)-(3), (7), and Section 2.4.3, Eq. (10)]
    "we employ a weighted expectation-maximization (EM) algorithm that integrates the physics score s_i as a sample-specific prior to update the initial binary label y_init. This process yields the final refined binary label ŷ_i, which is used for yielding cleaned visual features (m′_U,m′_M,m′_L)."

    The physics score s_i in Eq. (1) is built from n_e, T_e, f_G, and the EM Gaussian mixture uses x_i=[n_e,T_e,f_G,t] with s_i as prior (Eqs. 2-3). Eq. (7) then zeroes the visual features wherever ŷ_i=0. The same n_e, T_e, f_G are model inputs (Table 2) and form the physics gate g(f_G,T_e) in Eq. (10), whose thresholds come from the same Section 2.2.2 analysis. Consequently the refined features and the predicted label are constructed from the same physical variables the predictor consumes, so improved downstream metrics are consistent with a self-consistent loop rather than with an external MARFE benchmark.

full rationale

The central quantitative claim—40 ms-ahead MARFE worsening warning with AUC 0.969/F1 0.862—rests on the worsening label b_j(t) defined in Eq. (8). That label is a thresholded future increment of the cleaned visual features m′_j, and those same m′_j are explicit model inputs in Table 2; the thresholds θ_j are fit to training data. The evaluation therefore demonstrates that the model can anticipate a threshold crossing in its own cleaned input stream, which is a meaningful forecasting task but is not by itself evidence of detecting real MARFE onset. The label-generation loop is tightened by the EM refinement: the physics score s_i and Gaussian mixture use n_e, T_e, f_G, t, the same variables that enter the model inputs and the physics gate g(f_G,T_e), and Eq. (7) zeroes the input features wherever the EM says no MARFE. This makes the 'physics-informed' ground truth a function of the predictor's own input variables. The paper explicitly acknowledges in §3.1: 'Due to the lack of a comprehensive manually-labeled ground truth...', and its validation of the labels is qualitative (corrections cluster in the early discharge phase; representative shot #11522). The abstract/body discrepancies (AUC 0.981 vs 0.969, F1 0.840 vs 0.862, 140-shot vs ≈70-shot test split) further indicate protocol sensitivity, though these are correctness rather than circularity issues. No load-bearing self-citation chain is present, and the ODE model could still fail at the internal forecasting task, so the circularity is partial rather than total; the score is 6 rather than 8.

Assumptions & free parameters 8 free parameters · 5 assumptions · 0 invented entities

The central claim rests on a sequence of fitted thresholds and hand-chosen weights: physics prior thresholds from noisy labels, worsening-label jump thresholds from training data, gate gains, and EM hyperparameters. The paper adds no new physical entity, but the label construction is tightly coupled to the model inputs and to the physics variables used in the gate.

free parameters (8)
  • Physics prior thresholds (n_e mid/high, T_e mid/low, f_G mid/high) = 2.496e19, 3.765e19 m-3; 0.766, 0.668 keV; 0.741, 1.043
    Derived via Youden's J on ROC of the initial noisy labels and quantiles of the MARFE-positive distribution (Table 1). These thresholds define the physics score in Eq. (1).
  • Physics score weights in Eq. (1) = 0.2, 0.1, 0.2, 0.1, 0.3, 0.1
    Chosen by hand based on known physical importance; no sensitivity analysis or optimization criterion is stated.
  • Worsening jump threshold theta_j per zone = not reported
    Determined empirically from training data (Sec. 2.4.2) and used to define the ground-truth label in Eq. (8); values are not given in the paper.
  • Physics gate gains k_n and k_T = not reported
    Appear in Eq. (10) but no values or fitting procedure are given; the gate is the core physics-constraint contribution.
  • EM label threshold gamma_thre = 0.5
    Chosen to balance recall and precision (Sec. 2.3); no ablation on this threshold.
  • EM learning rate alpha = 0.5
    Tempered update in the M-step; fixed hyperparameter with no sensitivity analysis.
  • Jump multiplier c = 1.5
    Arbitrary constant in Eq. (8) separating notable from exceptionally large jumps.
  • Brightness threshold T_bright and opening kernel = 220, 5x5
    Image-processing constants used to derive initial features; fixed without calibration against expert labels.
assumptions (5)
  • domain assumption MARFE is a radiative thermal instability on the high-field side and is linked to the Greenwald density limit.
    Used to justify the physics prior, gate, and label cleaning; cited to Lipschultz and Greenwald, but not independently verified on HL-3 within this paper.
  • domain assumption Core electron density n_e can stand in for line-averaged density when computing f_G.
    Acknowledged in Sec. 2.2.2 as a diagnostic-availability choice; the paper asserts trend robustness without quantitative validation.
  • domain assumption Visual CCD features after thresholding and morphological opening correspond to MARFE morphology.
    No expert-labeled ground truth or camera calibration is provided; the EM refinement is the only correction.
  • standard math EM with diagonal GMM converges to a meaningful separation of MARFE and non-MARFE states.
    Standard EM theory is assumed; convergence diagnostics and initialization sensitivity are not reported.
  • standard math Neural ODE training via adjoint sensitivity is stable for this task.
    Assumed standard; no solver error tolerance or numerical stability analysis is reported.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Physics-Gated Visual Prediction of MARFE on the HL-3 Tokamak." pith.science (2026). https://pith.science/paper/6DJC23NC

@misc{pith2026251024347,
  author       = {Pith},
  title        = {Pith review of: Physics-Gated Visual Prediction of MARFE on the HL-3 Tokamak},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/6DJC23NC}},
  note         = {Machine review of arXiv:2510.24347}
}
abstract

The Multifaceted Asymmetric Radiation From the Edge (MARFE) is a critical plasma instability that often precedes density-limit disruptions in tokamaks, posing a significant risk to machine integrity and operational efficiency. We develop a physics-gated, continuous MARFE monitor for the HL-3 tokamak that outputs a per-frame intensity probability every $2$\,ms, which can potentially be used by the shape-target controller in the plasma control system. Our framework integrates two core innovations: (1) a physics-scored, weighted Expectation-Maximization (EM) pipeline that refines noisy visual labels using $(n_e, T_e, f_G, t)$ as a Bayesian prior, and (2) a continuous-time, physics-gated Neural Ordinary Differential Equation (Neural ODE) backbone whose dynamics are modulated by a sigmoid gate on $f_G$ and $T_e$. Meanwhile, the Neural ODE adopts a $40$\,ms forward forecasting horizon to accommodate the actuator-response budget. On a frozen $140$-shot held-out test set, the proposed method yields a median label-aligned lead time of $+36$\,ms, close to this design horizon. Against a Bi-LSTM baseline trained under the matched protocol, the proposed Neural ODE attains Area Under the Curve (AUC) $=0.981$ and sample-level $F_1=0.840$, compared with AUC $=0.960$ and sample-level $F_1=0.779$ for the baseline. The deployed inference service runs within a $1$-ms control-cycle budget, while new diagnostic samples are generated at the $2$-ms frame cadence.

Review history (2 revisions) →

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

40 extracted references · 34 canonical work pages

  1. [1]

    MARFE worsening

    TC Hender, JC Wesley, J Bialek, A Bondeson, AH Boozer, RJ Buttery, A Garofalo, TP Goodman, RS Granetz, Y Gribov, Physics-Informed Visual Prediction of MARFE on the HL-3 Tokamak12 Table 5: Nomenclature and List of Symbols. Symbol Description General Plasma Parameters ne Core electron density. fG Normalized density (ne/ng). Te Core electron temperature. q95...

    work page 2007

  2. [2]

    Review of MARFE phenomena in tokamaks

    Bruce Lipschultz. Review of MARFE phenomena in tokamaks. Journal of Nuclear Materials, 145:15–25, 1987

    work page 1987

  3. [3]

    MARFE: An edge plasma phenomenon.Nuclear Fusion, 24(8):977, 1984

    B Lipschultz, B LaBombard, ES Marmar, MM Pickrell, JL Terry, R Watterson, and SM Wolfe. MARFE: An edge plasma phenomenon.Nuclear Fusion, 24(8):977, 1984

    work page 1984

  4. [4]

    A new look at density limits in tokamaks.Nuclear Fusion, 28(12):2199, 1988

    Martin Greenwald, JL Terry, SM Wolfe, S Ejima, MG Bell, SM Kaye, and GH Neilson. A new look at density limits in tokamaks.Nuclear Fusion, 28(12):2199, 1988

    work page 1988

  5. [5]

    FA Kelly, WM Stacey, J Rapp, and M Brix. Thermal instability theory analysis of multifaceted asymmetric radiation from the edge (MARFE) in tokamak experiment for technology oriented research (TEXTOR).Physics of Plasmas, 8(7):3382–3390, 2001

    work page 2001

  6. [6]

    U. Samm, M. Brix, F. Durodi´e, M. Lehnen, A. Pospieszczyk, J. Rapp, G. Sergienko, B. Schweer, M. Z. Tokar, and B. Unterberg. MARFE feedback experiments on TEXTOR-94.Journal of Nuclear Materials, 266-269:666–672, 1999

    work page 1999

  7. [7]

    The ITER design

    R Aymar, P Barabaschi, and Y Shimomura. The ITER design. Plasma Physics and Controlled Fusion, 44(5):519, 2002

    work page 2002

  8. [8]

    Iter construction and manufacturing progress toward first plasma.Fusion Engineering and Design, 146:124–129, 2019

    Bernard Bigot. Iter construction and manufacturing progress toward first plasma.Fusion Engineering and Design, 146:124–129, 2019

    work page 2019

Show all 40 references
  1. [9]

    Status of research toward the ITER disruption mitigation system.Physics of Plasmas, 22(2), 2015

    EM Hollmann, PB Aleynikov, T ¨unde F ¨ul¨op, DA Humphreys, V A Izzo, M Lehnen, VE Lukash, Gergely Papp, G Pautasso, F Saint-Laurent, et al. Status of research toward the ITER disruption mitigation system.Physics of Plasmas, 22(2), 2015

    2015

  2. [10]

    Improved radiation measurements on JET–first results from an upgraded bolometer system.Journal of Nuclear Materials, 363:365–370, 2007

    A Huber, K McCormick, P Andrew, MR de Baar, P Beaumont, S Dalley, J Fink, JC Fuchs, K Fullard, W Fundamenski, et al. Improved radiation measurements on JET–first results from an upgraded bolometer system.Journal of Nuclear Materials, 363:365–370, 2007

    2007

  3. [11]

    Application of axuv diode detectors at ASDEX upgrade.Review of scientific Instruments, 85(3), 2014

    M Bernert, T Eich, A Burckhart, JC Fuchs, L Giannone, A Kallenbach, RM McDermott, B Sieglin, ASDEX Upgrade Team, et al. Application of axuv diode detectors at ASDEX upgrade.Review of scientific Instruments, 85(3), 2014

    2014

  4. [12]

    A control oriented strategy of disruption prediction to avoid the configuration collapse of tokamak reactors.Nature Communications, 15(1):2424, 2024

    Andrea Murari, Riccardo Rossi, Teddy Craciunescu, Jes ´us Vega, and Michela Gelfusa. A control oriented strategy of disruption prediction to avoid the configuration collapse of tokamak reactors.Nature Communications, 15(1):2424, 2024

    2024

  5. [13]

    Algorithms for the automatic identification of MARFEs and UFOs in JET database of visible camera videos.IEEE Transactions on Plasma Science, 38(12):3409–3418, 2010

    Andrea Murari, Massimo Camplani, Barbara Cannas, D Mazon, F Delaunay, P Usai, and JF Delmond. Algorithms for the automatic identification of MARFEs and UFOs in JET database of visible camera videos.IEEE Transactions on Plasma Science, 38(12):3409–3418, 2010

    2010

  6. [14]

    Phase congruency image classification for MARFE detection on JET with a carbon wall.Fusion Science and Technology, 62(2):339–346, 2012

    T Craciunescu, A Murari, I Tiseanu, J Vega, and JET-EFDA Contributors. Phase congruency image classification for MARFE detection on JET with a carbon wall.Fusion Science and Technology, 62(2):339–346, 2012

    2012

  7. [15]

    Portes de Albuquerque, M

    M. Portes de Albuquerque, M. P. de Albuquerque, G. Chacon, E. L. de Faria, A. Murari, and JET EFDA contributors. High-speed image processing algorithms for real-time detection of MARFEs on JET.IEEE Transactions on Plasma Science, 40(12):3485– 3492, 2012

    2012

  8. [16]

    Gonz ´alez Ganz´abal, G

    A. Gonz ´alez Ganz´abal, G. A. Ratt ´a, D. Gadariya, and S. Dormido- Canto. Advancing MARFE detection in JET’s operational camera videos through machine learning techniques.Fusion Engineering and Design, 205:114534, 2024

    2024

  9. [17]

    Wenhui Hu, Jilei Hou, Zhengping Luo, Yao Huang, Dalong Chen, Bingjia Xiao, Qiping Yuan, Yanmin Duan, Jiansheng Hu, Guizhong Zuo, et al. Prediction of multifaceted asymmetric radiation from the edge movement in density-limit disruptive plasmas on experimental advanced supercond...

    2023

  10. [18]

    Predicting disruptive instabilities in controlled fusion plasmas through deep learning.Nature, 568(7753):526–531, 2019

    Julian Kates-Harbeck, Alexey Svyatkovskiy, and William Tang. Predicting disruptive instabilities in controlled fusion plasmas through deep learning.Nature, 568(7753):526–531, 2019

    2019

  11. [19]

    CI Stuart, G Artaserse, P Card, IS Carvalho, R Felton, SN Gerasi- mov, A Goodyear, RB Henriques, D Karkinsky, PJ Lomas, et al. Petra: A generalised real-time event detection platform at JET Physics-Informed Visual Prediction of MARFE on the HL-3 Tokamak13 for disruption predic...

    2021

  12. [20]

    Observations with fast visible cameras in high power deuterium plasma experiments in the JET ITER-like wall tokamak.Nuclear Materials and Energy, 25:100837, 2020

    Ulises Losada, A Manzanares, I Balboa, S Silburn, J Karhunen, Pedro J Carvalho, A Huber, V Huber, Emilia R Solano, E De La Cal, et al. Observations with fast visible cameras in high power deuterium plasma experiments in the JET ITER-like wall tokamak.Nuclear Materials and Ener...

    2020

  13. [21]

    Wall reflection issues for optical diagnostics in fusion devices.Review of Scientific Instruments, 81(10), 2010

    Ph Lotte, MH Aumeunier, P Devynck, C Fenzi, V Martin, and JM Trav `ere. Wall reflection issues for optical diagnostics in fusion devices.Review of Scientific Instruments, 81(10), 2010

    2010

  14. [22]

    M. Carr, A. Meakins, S. A. Silburn, J. Karhunen, M. Bernert, and et al. Physically principled reflection models applied to filtered camera imaging inversions in metal walled fusion machines. Review of Scientific Instruments, 90(4):043504, 2019

    2019

  15. [23]

    The fast camera (fastcam) imaging diagnostic systems on the DIII-D tokamak.Review of Scientific Instruments, 94(5), 2023

    Claudio Marini, JA Boedo, EM Hollmann, L Chousal, J Mills, Z Popovi ´c, and I Bykov. The fast camera (fastcam) imaging diagnostic systems on the DIII-D tokamak.Review of Scientific Instruments, 94(5), 2023

    2023

  16. [24]

    Maximum likelihood from incomplete data via the EM algorithm.Journal of the royal statistical society: series B (methodological), 39(1):1– 22, 1977

    Arthur P Dempster, Nan M Laird, and Donald B Rubin. Maximum likelihood from incomplete data via the EM algorithm.Journal of the royal statistical society: series B (methodological), 39(1):1– 22, 1977

    1977

  17. [25]

    John Wiley & Sons, 2008

    Geoffrey J McLachlan and Thriyambakam Krishnan.The EM algorithm and extensions. John Wiley & Sons, 2008

    2008

  18. [26]

    A comprehensive study of the uncertainties in bolometric tomography on JET using the maximum likelihood method.Review of Scientific Instruments, 90(12), 2019

    E Peluso, T Craciunescu, A Murari, P Carvalho, M Gelfusa, and Contributors JET. A comprehensive study of the uncertainties in bolometric tomography on JET using the maximum likelihood method.Review of Scientific Instruments, 90(12), 2019

    2019

  19. [27]

    Maximum likelihood bolometry for ASDEX upgrade experiments.Physica Scripta, 98(12):125603, 2023

    Teddy Craciunescu, Emmanuele Peluso, Andrea Murari, Matthias Bernert, Michela Gelfusa, Riccardo Rossi, Luca Spolladore, Ivan Wyss, Pierre David, Stuart Henderson, et al. Maximum likelihood bolometry for ASDEX upgrade experiments.Physica Scripta, 98(12):125603, 2023

    2023

  20. [28]

    Raissi, P

    M. Raissi, P. Perdikaris, and G. E. Karniadakis. Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations.Journal of Computational Physics, 378:686–707, 2019

    2019

  21. [29]

    Integrating physics-based modeling with machine learning: A survey.arXiv preprint arXiv:2003.04919, 1(1):1–34, 2020

    Jared Willard, Xiaowei Jia, Shaoming Xu, Michael Steinbach, and Vipin Kumar. Integrating physics-based modeling with machine learning: A survey.arXiv preprint arXiv:2003.04919, 1(1):1–34, 2020

    2003 arXiv

  22. [30]

    Neural ordinary differential equations

    Ricky TQ Chen, Yulia Rubanova, Jesse Bettencourt, and David K Duvenaud. Neural ordinary differential equations. InAdvances in Neural Information Processing Systems, volume 31, Montr´eal, Canada, 2018

    2018

  23. [31]

    Latent ordinary differential equations for irregularly-sampled time series

    Yulia Rubanova, Ricky TQ Chen, and David K Duvenaud. Latent ordinary differential equations for irregularly-sampled time series. InAdvances in Neural Information Processing Systems, volume 32, Vancouver, Canada, 2019

    2019

  24. [32]

    Neural controlled differential equations for irregular time series

    Patrick Kidger, James Morrill, James Foster, and Terry Lyons. Neural controlled differential equations for irregular time series. InAdvances in Neural Information Processing Systems, volume 33, Virtual, 2020

    2020

  25. [33]

    Progress of hl-2a experiments and hl-2m program.Nuclear Fusion, 62(4):042020, 2022

    XR Duan, M Xu, WL Zhong, Y Liu, XM Song, DQ Liu, YQ Wang, B Lu, ZB Shi, GY Zheng, et al. Progress of hl-2a experiments and hl-2m program.Nuclear Fusion, 62(4):042020, 2022

    2022

  26. [34]

    Index for rating diagnostic tests.Cancer, 3(1):32–35, 1950

    William J Youden. Index for rating diagnostic tests.Cancer, 3(1):32–35, 1950

    1950

  27. [35]

    Model for access and stability of the x-point radiator and the threshold for MARFES in tokamak plasmas.Nuclear Fusion, 62(7):076008, 2022

    U Stroth, M Bernert, D Brida, M Cavedon, R Dux, E Huett, T Lunt, O Pan, M Wischmeier, ASDEX Upgrade Team, et al. Model for access and stability of the x-point radiator and the threshold for MARFES in tokamak plasmas.Nuclear Fusion, 62(7):076008, 2022

    2022

  28. [36]

    Accurately computing the log-sum-exp and softmax functions

    Pierre Blanchard, Desmond J Higham, and Nicholas J Higham. Accurately computing the log-sum-exp and softmax functions. IMA Journal of Numerical Analysis, 41(4):2311–2330, 2021

    2021

  29. [37]

    Adams and David J

    Ryan P. Adams and David J. C. MacKay. Bayesian online changepoint detection.arXiv preprint arXiv:0710.3742, 2007

    2007 arXiv

  30. [38]

    Gru-ode-bayes: Continuous modeling of sporadically-observed time series.Advances in Neural Information Processing Systems, 32, 2019

    Edward De Brouwer, Jaak Simm, Adam Arany, and Yves Moreau. Gru-ode-bayes: Continuous modeling of sporadically-observed time series.Advances in Neural Information Processing Systems, 32, 2019

    2019

  31. [39]

    Multi-task learning using uncertainty to weigh losses for scene geometry and semantics

    Alex Kendall, Yarin Gal, and Roberto Cipolla. Multi-task learning using uncertainty to weigh losses for scene geometry and semantics. InProceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pages 7482–7491, Salt Lake City, UT, USA, 2018

    2018

  32. [40]

    Hindmarsh

    Radu Serban and Alan C. Hindmarsh. Cvodes, the sensitivity- enabled ode solver in sundials. InProceedings of IDETC/CIE 2005, ASME International Design Engineering Technical Confer- ences, Long Beach, CA, USA, 2005. Adjoint module uses check- pointing and Hermite interpolation

    2005

Pith tools

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