Pith. sign in

REVIEW 5 major objections 8 minor 32 references

Towards energy-insensitive and robust neutron/gamma classification: A learning-based frequency-domain parametric approach

T0 review · 5 major / 8 minor · reviewed 2026-08-08 · deepseek-v4-flash

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

Pith's one-line read A minimal feedforward network on the FFT spectrum, trained on charge-normalized pulses, outperforms fixed discriminators for neutron/gamma separation, especially at low sampling rates and under added noise.

desk verdict A simple FFT-based learned classifier with charge normalization that plausibly beats FCI at low sampling rates, wrapped in a useful adversarial-ROC evaluation—not a result that changes physics, but one worth referee time. read the letter →

arxiv 2502.07359 v3 pith:3K46WDYE submitted 2025-02-11 physics.ins-det physics.data-an

Pengcheng Ai , Hongtao Qin , Xiangming Sun , Kaiwen Shang This is my paper · ORCID
classification physics.ins-detphysics.data-an
keywords neutron/gammadiscriminationpulseshapefrequency-domainparametricmodelschargenormalizationadversarialsamplingCLYCscintillatorROCcurvefigure-of-merit
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper tries to establish that simple frequency-domain parametric models (FDPMs)—basic feedforward networks fed with the FFT spectrum of a pulse—are better neutron/gamma discriminators than fixed indexes like the frequency classification index (FCI), once the data are charge-normalized. It argues that figure-of-merit (FoM) is too coarse for high-precision discrimination and proposes evaluating classifiers with ROC curves on adversarially selected events that sit near the classification boundary. On an open-source CLYC dataset, the nonlinear FDPM reaches near-perfect adversarial AUC at every tested sampling rate, including 10 MS/s where FCI drops to 0.751, and it remains the best method under added Gaussian noise. If this holds, cheap, low-rate digitizers with simple neural-network inference could replace high-rate fixed-index discrimination.

What carries the argument

The machinery is the FDPM: a full N-point FFT of each baseline-subtracted pulse, transformed to real features by $f^r = |\mathrm{Re}(f^c)| + |\mathrm{Im}(f^c)|$, then either a linear map $p_l = W^l f^r + b^l$ or a one-hidden-layer nonlinear map $p_n = W^n_2 \sigma(W^n_1 f^r + b^n_1) + b^n_2$, trained with softmax cross-entropy. Charge normalization (scaling every pulse to the same integral) is the step that forces the model to learn energy-independent features. The evaluation machinery is adversarial sampling: selecting neutron events with the smallest amplitudes and gamma events with the largest amplitudes so that the two classes overlap in energy, making ROC curves informative where FoM cannot distinguish methods.

What would settle it

Train the nonlinear FDPM on charge-normalized events split into narrow energy bins, then measure AUC within each bin; if AUC varies strongly across energy or drops well below 0.999 in some bins, the energy-insensitivity claim fails. Alternatively, generate synthetic pulses with identical shapes but different integrals, normalize them, and check whether the network still separates them—if it does, charge normalization leaks energy information.

Watch

Extended reading notes

Core claim

The central discovery is that a learned combination of the full discrete FFT spectrum separates neutrons from gammas using energy-independent pulse-shape information, not just the energy gap. A linear model (one matrix) already beats FCI at 20 and 10 MS/s, and a nonlinear model with one hidden layer reaches adversarial AUC 1.000 at 100 and 20 MS/s and 0.999 at 10 MS/s, while FCI falls to 0.751. Charge normalization is essential: without it, the linear model's ROC collapses to the energy-only baseline (AUC 0.675–0.684), showing that the network would otherwise exploit the neutron/gamma energy gap. The same approach transfers to a TOF-labelled organic scintillator dataset, where FDPM and CNN give similar ROC curves and the best conventional method is slightly better on the AUC metric.

Load-bearing premise

The claim that the classifier is energy-insensitive rests on charge normalization removing energy information without distorting neutron/gamma pulse-shape differences; if normalization instead encodes class-dependent artifacts, the reported near-perfect separation is a preprocessing effect rather than true pulse-shape discrimination.

Editorial extensions

If this is right

  • At 10 MS/s, the nonlinear FDPM keeps adversarial AUC 0.999 and 98.5% accuracy, while FCI falls to 0.751 and 61.5%, so low-rate digitizers should no longer be ruled out for high-precision neutron/gamma discrimination.
  • Under 30 dB Gaussian noise, the nonlinear FDPM still leads across sampling rates, and it degrades much less with correlated noise and dark pulses than with white noise.
  • The linear FDPM's learned frequency weights are interpretable and concentrate in the low-frequency region, matching the physical intuition behind fixed frequency indexes while being optimized for the data.
  • Because the linear FDPM at 10 MS/s needs only 256-point FFTs and matrix operations, the method is suitable for FPGA or microcontroller implementation and online inference.

Reading between the lines

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

  • A natural next test is to apply FDPM to pile-up events or lower-bit ADC data; the frequency-domain features may degrade differently than time-domain charge comparison, and the current paper does not test these regimes.
  • The charge-normalization trick could be replaced by other energy-invariant transformations, such as normalizing by peak amplitude or by a low-frequency bin, and comparing those would isolate whether the energy insensitivity comes from normalization or from the FFT features themselves.
  • If the method generalizes, learned frequency weights could serve as a physically interpretable replacement for manually tuned discriminator indexes, and the same adversarial-sampling ROC protocol could be adopted as a standard benchmark for high-precision neutron/gamma classifiers.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

5 major / 8 minor

Summary. The paper proposes linear and nonlinear frequency-domain parametric models (FDPM) for neutron/gamma discrimination. The models take FFT-derived features from charge-normalized pulses and learn a small set of weights via cross-entropy training. The authors argue that figure-of-merit (FoM) is inadequate for high-precision discriminators and introduce ROC curves on adversarially sampled events as a more stringent evaluation. Using an open-source CLYC dataset, they report that the nonlinear FDPM achieves near-perfect AUC on adversarial examples at sampling rates from 100 MS/s down to 10 MS/s (AUC 0.999–1.000), outperforming the frequency classification index (FCI), which degrades at low rates, as well as conventional methods. Additional experiments address noise robustness, model interpretability, and an organic scintillator dataset.

Significance. If the results are reproducible, the paper offers a simple, interpretable, and hardware-friendly classifier that improves neutron/gamma discrimination on difficult events, and it proposes an evaluation criterion that is more informative than FoM for high-precision applications. Strengths include the use of an open dataset, explicit ablations (w/o FFT, w/o normalization), and a clear statement that charge normalization is essential for energy insensitivity. However, several load-bearing aspects of the evaluation protocol need clarification or strengthening before the central claims can be fully accepted.

major comments (5)
  1. [Sec. 4.2, FCI description] The FCI baseline has its cut-off frequencies 'adjusted for best performance according to the sampling rate,' but the tuning procedure (search range, optimization criterion, use of validation or test set) is not described. Because FDPM parameters are learned on a fixed training split, the comparison is not on equal footing unless FCI is tuned on the same training/validation split only. Please specify the tuning protocol and report the chosen cut-off values and their sensitivity.
  2. [Table 1 and Sec. 4.2/4.3] The Energy baseline reports exactly 53 errors (both on all data and on adversarial examples) and AUC 0.685 for every sampling rate and every noise condition. This is suspicious: if the Energy discriminator is applied to subsampled or noisy waveforms, some variation in error counts would be expected. Please clarify whether the Energy baseline is computed on the original full-rate, noiseless waveforms for all conditions. Also state the number of adversarial examples (the error counts imply 200) and whether the same adversarial set is reused across all conditions; report variability by repeating the adversarial selection and by training multiple models with different seeds.
  3. [Sec. 4.1, preprocessing steps 3–4, and Sec. 4.2/4.3] It is unclear whether a separate model is trained for each sampling rate and each noise condition, or whether a single trained model is evaluated under all conditions. The input dimension N changes with subsampling, so separate models appear necessary, but the text does not state this. Please specify the training protocol for every row of Table 1, including whether noise is added before the train/test split or only to the test set. This is central to the claim of 'adaptability to variations of data integrity.'
  4. [Sec. 4.5, Table 3] Conventional methods (charge comparison, zero crossing, PCNN, ladder gradient, FGA, FEPS) are evaluated with min-max normalization, while FDPM and CNN/LSTM are evaluated with charge normalization. This difference in preprocessing could disadvantage the conventional methods. Please justify the choice of min-max normalization for each conventional method, or evaluate them under the same charge normalization when their definitions permit, and report the sensitivity of the results to the normalization choice.
  5. [Sec. 2.2 and Sec. 4.1] The adversarial sampling algorithm is underspecified. The text says examples are 'shared by several discrimination methods' and that for CLYC, neutron events with smallest amplitudes and gamma events with largest amplitudes are selected, but the exact number of events, the threshold criteria, and which methods contribute to the 'shared' condition are not stated. It is also not explicit that adversarial examples are selected only from the test set. Please provide the full algorithm and confirm that no information from the test set leaks into training or validation.
minor comments (8)
  1. [Eq. (5)] The term 'city-block approach' for |Re(f^c)| + |Im(f^c)| is nonstandard; please define it clearly and justify why this combination is used instead of the complex magnitude.
  2. [Sec. 4.1, step 5] Please specify the target integral value used for charge normalization and how zero-integral or near-zero-integral signals are handled.
  3. [Sec. 4.2, discussion of Fig. 4] At 50 MS/s the nonlinear FDPM has AUC 0.965 versus FCI 0.996, so the statement that FDPM models 'significantly improve' performance is too strong; consider wording such as 'improve robustness across sampling rates.'
  4. [Sec. 4.3, Fig. 5] The legend of Fig. 5 is crowded with many curves; consider splitting into separate panels or presenting the AUC values in a table for readability.
  5. [Sec. 4.4, Fig. 6] Please define how 'noise level' is computed in Fig. 6a and 6b (e.g., standard deviation divided by maximum amplitude) and explicitly relate the dB SNR values to the plotted noise levels.
  6. [Sec. 4.6, LSTM results] The paper notes that LSTM converged in only 2 of 5 folds on the organic dataset; please state this limitation in the table caption and consider excluding the non-converged folds from the reported metrics.
  7. [Sec. 1, contributions list] Typo: 'base on FDPM' should be 'based on FDPM.' Also, 'the larger value of the two targets will be the predicted class' requires a comma or semicolon for readability.
  8. [References] For the dataset reference [23], please include the version or access date, as Zenodo datasets may be updated.

Circularity Check

1 steps flagged · score 2.0 of 10

No significant circularity: energy-insensitivity is enforced by charge normalization rather than discovered, but the paper discloses and ablates this, and the core accuracy comparisons are held-out empirical results.

  1. self definitional [Section 4.1, preprocessing step 5; discussed in Section 4.2]
    "Finally, to force models to learn energy-independent features from the signal, all examples are normalized to have the same integral values. It should be noted that the final step is essential for models to be energy-insensitive. This will be shown in the subsection below."

    The 'energy-insensitive' property is not discovered from the data; it is enforced by construction, because normalizing every pulse to the same integral removes the total-charge/energy feature before the model sees the input. Therefore any classifier trained on this preprocessing is necessarily insensitive to total charge, making the later claim that FDPM is energy-insensitive partly a restatement of the preprocessing rather than an independently derived property. The paper is transparent about this, explicitly says the step is 'essential', and supports it with an ablation (Linear w/o norm. AUC 0.675 vs Energy 0.685 in Fig. 4) showing that without normalization the linear model reduces to essentially an energy discriminator.

full rationale

The central empirical claims—FDPM outperforms FCI and other methods at low sampling rates and under noise—are supported by held-out test data: models are trained on a 7:1:2 split and evaluated on the test portion, with adversarial examples drawn from the test set. There is no fitted parameter renamed as a prediction, and no result is derived from the conclusion being tested. The paper's only mildly circular element is the energy-insensitivity claim, which is guaranteed by charge normalization: equalizing integrals by definition removes total energy as an input feature. However, the authors state this intention explicitly, call the step 'essential', and verify the consequence with an ablation and with adversarial sampling that equalizes energy across classes. Consequently, the energy-insensitivity statement is an honest description of the preprocessing rather than a concealed equivalence. The comparison is not perfectly even—FCI cut-off frequencies are hand-tuned per sampling rate while FDPM is trained once, and the reused adversarial example set shows suspiciously constant Energy-baseline error counts—but these are experimental-design and fairness concerns, not circular reasoning. No load-bearing self-citation chain, imported uniqueness theorem, or ansatz-smuggling-via-citation occurs; the FFT-based starting point is explicitly credited to prior non-overlapping work. Overall, the paper's derivation chain is self-contained and its main accuracy claims stand on independent empirical evaluation.

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

No new physical entities are introduced. The central empirical claims rest on labeled data, a normalization scheme, and a self-defined evaluation subset; these are the main assumptions to scrutinize.

free parameters (4)
  • FDPM network weights and biases (W_l, b_l, W_n1, b_n1, W_n2, b_n2) = not reported; model checkpoints not released
    Optimized on the CLYC training set by cross-entropy loss; these fitted parameters constitute the classifier whose performance is claimed.
  • FCI cutoff frequencies = not stated numerically
    Adjusted 'for best performance according to the sampling rate' (Section 4.2); hand-tuned per condition, affecting the FCI baseline ROC/AUC.
  • Adversarial sample size and selection thresholds = 200 CLYC examples; 1000 organic examples
    The authors choose how many and which examples form the hard evaluation subset (low-energy neutrons versus high-energy gammas for CLYC; near charge-comparison crossing for organic). AUC and error counts depend on this choice.
  • Training hyperparameters = lr=0.001, momentum=0.9, 20 epochs, batch=512, hidden size N/2
    Chosen by hand without reported sensitivity analysis; could affect the comparison.
assumptions (4)
  • domain assumption Charge normalization preserves the pulse-shape information needed for neutron/gamma discrimination while removing energy.
    Invoked in Section 4.1 step 5 and Section 4.2; if normalization distorts shapes or introduces class-dependent artifacts, the model's energy-insensitivity is an artifact.
  • domain assumption Ground-truth labels in the CLYC dataset [23] correctly tag neutron and gamma events.
    Supervised training and all reported metrics assume the source-based labels are accurate; mis-tagged events would bias every method and the adversarial subset.
  • ad hoc to paper Adversarial sampling on energy (CLYC) or charge-comparison PSD (organic) is a valid criterion for comparing high-precision discriminators.
    Proposed in Section 2.2 as the paper's main evaluation criterion; no independent validation that this subset predicts real-world field performance.
  • domain assumption The FFT city-block magnitude representation discards phase without losing class-discriminative information.
    Equation (5) uses |Re(f_c)| + |Im(f_c)|; the paper does not compare against phase-sensitive features.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Towards energy-insensitive and robust neutron/gamma classification: A learning-based frequency-domain parametric approach." pith.science (2026). https://pith.science/paper/3K46WDYE

@misc{pith2026250207359,
  author       = {Pith},
  title        = {Pith review of: Towards energy-insensitive and robust neutron/gamma classification: A learning-based frequency-domain parametric approach},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/3K46WDYE}},
  note         = {Machine review of arXiv:2502.07359}
}
read the original abstract

Neutron/gamma discrimination has been intensively researched in recent years, due to its unique scientific value and widespread applications. With the advancement of detection materials and algorithms, nowadays we can achieve fairly good discrimination. However, further improvements rely on better utilization of detector raw signals, especially energy-independent pulse characteristics. We begin by discussing why figure-of-merit (FoM) is not a comprehensive criterion for high-precision neutron/gamma discriminators, and proposing a new evaluation method based on adversarial sampling. Inspired by frequency-domain analysis in existing literature, parametric linear/nonlinear models with minimum complexity are created, upon the discrete spectrum, with tunable parameters just as neural networks. We train the models on an open-source neutron/gamma dataset (CLYC crystals with silicon photomultipliers) preprocessed by charge normalization to discover and exploit energy-independent features. The performance is evaluated on different sampling rates and noise levels, in comparison with the frequency classification index and conventional methods. The frequency-domain parametric models show higher accuracy and better adaptability to variations of data integrity than other discriminators. The proposed method is also promising for online inference on economical hardware and portable devices.

Figures

Figures reproduced from arXiv: 2502.07359 by the authors.

Figure 1
Figure 1. These sub-figures illustrate that distributions with different numbers of misclassified examples can share the same [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Illustration of 100 neutron examples and 100 gamma examples with (a) adversarial sampling and (b) random sampling. [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Illustration of the proposed FDPM. The parameters are tuned offline and deployed online for inference. [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (7 more)
Figure 4
Figure 4. Figure 4: ROC curves by different discrimination methods at sampling rates of 100 MS/s, 50 MS/s, 20 MS/s and 10 MS/s. [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]
Figure 5
Figure 5. Figure 5: ROC curves by different discrimination methods at sampling rates of 100 MS/s, 50 MS/s, 20 MS/s and 10 MS/s, [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]
Figure 6
Figure 6. Figure 6: The performance sensitivity of the proposed method on different noise components. (a) Gaussian white noise. (b) [PITH_FULL_IMAGE:figures/full_fig_p010_6.png]
Figure 7
Figure 7. Figure 7: ROC curves by different discrimination methods at the sampling rate of 10 MS/s. The numbers in the parentheses [PITH_FULL_IMAGE:figures/full_fig_p012_7.png]
Figure 8
Figure 8. Figure 8: Histograms illustrating the distributions of misclassified examples over energy/height. The distributions of all test [PITH_FULL_IMAGE:figures/full_fig_p012_8.png]
Figure 9
Figure 9. Figure 9: Additional ROC curves by different discrimination methods for the organic dataset. The numbers in the parentheses [PITH_FULL_IMAGE:figures/full_fig_p013_9.png]
Figure 10
Figure 10. Figure 10: Visualizing weights of the linear FDPM on the discrete spectrum. (a) CLYC dataset with the 10 MS/s sampling [PITH_FULL_IMAGE:figures/full_fig_p014_10.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

32 extracted references · 26 canonical work pages

  1. [1]

    H. A. Imane Ahnouz, R. Sebihi, A Review of Neutron–Gamma-Ray Discrimination Methods Using Organic Scintillators, Nuclear Science and Engineering 198 (12) (2024) 2241–2273. doi:10.1080/00295639.2024.2316946

    arXiv 2024

  2. [2]

    S¨ oderstr¨ om, G

    P.-A. S¨ oderstr¨ om, G. Jaworski, J. Valiente Dob´ on, J. Nyberg, J. Agramunt, G. de Angelis, S. Carturan, J. Egea, M. Er- duran, S. Ert¨ urk, G. de France, A. Gadea, A. Goasduff, V. Gonz´ alez, K. Hady´ nska-Kl¸ ek, T. H¨ uy¨ uk, V. Modamio, M. Moszynski, A. Di Nitto, M. Palacz, N. Pietralla, E. Sanchis, D. Testov, A. Triossi, R. Wadsworth, Neutron dete...

    work page 2019

  3. [3]

    S. Lee, K. Ko, G. Song, W. Kim, S. Yoon, C. Lee, K. T. Lim, G. Cho, Investigation of neutron/gamma-ray distribution in SiPM-based pulse shape discrimination using EJ-276 plastic scintillators, Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment 1066 (2024) 169638. doi:10.1016/ j.ni...

    work page arXiv 2024

  4. [4]

    Wolski, M

    D. Wolski, M. Moszy´ nski, T. Ludziejewski, A. Johnson, W. Klamra, ¨O. Skeppstedt, Comparison of n- γ discrimination by zero-crossing and digital charge comparison methods, Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment 360 (3) (1995) 584–592. doi:10.1016/0168-9002(95) 00037-2

    doi:10.1016/0168-9002(95 1995

  5. [5]

    Bayat, N

    E. Bayat, N. Divani-Vais, M. Firoozabadi, N. Ghal-Eh, A comparative study on neutron-gamma discrimination with NE213 and UGLLT scintillators using zero-crossing method, Radiation Physics and Chemistry 81 (3) (2012) 217–220. doi:10.1016/j.radphyschem.2011.10.016

    work page doi:10.1016/j.radphyschem.2011.10.016 2012

  6. [6]

    G. Liu, M. J. Joyce, X. Ma, M. D. Aspinall, A digital method for the discrimination of neutrons and γ rays with organic scintillation detectors using frequency gradient analysis, IEEE Transactions on Nuclear Science 57 (3) (2010) 1682–1691. doi:10.1109/TNS.2010.2044246

    arXiv 2010

  7. [7]

    Liu, B.-Q

    M.-Z. Liu, B.-Q. Liu, Z. Zuo, L. Wang, G.-B. Zan, X.-G. Tuo, Toward a fractal spectrum approach for neutron and gamma pulse shape discrimination, Chinese Physics C 40 (6) (2016) 066201. doi:10.1088/1674-1137/40/6/066201

    work page doi:10.1088/1674-1137/40/6/066201 2016

  8. [8]

    H. Ye, L. Chen, X. Xu, G. Jin, Fast FPGA algorithm for neutron–gamma discrimination, Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment 1027 (2022) 166256. doi:10.1016/j.nima.2021.166256

    work page arXiv 2022

Show all 32 references
  1. [9]

    Liu, Y.-X

    H.-R. Liu, Y.-X. Cheng, Z. Zuo, T.-T. Sun, K.-M. Wang, Discrimination of neutrons and gamma rays in plastic scintillator based on pulse-coupled neural network, Nuclear Science and Techniques 32 (8) (2021) 82. doi:10.1007/ s41365-021-00915-w . 15

    2021

  2. [10]

    Liu, M.-Z

    H.-R. Liu, M.-Z. Liu, Y.-L. Xiao, P. Li, Z. Zuo, Y.-H. Zhan, Discrimination of neutron and gamma ray using the ladder gradient method and analysis of filter adaptability, Nuclear Science and Techniques 33 (12) (2022) 159. doi:10.1007/ s41365-022-01136-5

    2022

  3. [11]

    Liu, H.-R

    B.-Q. Liu, H.-R. Liu, L. Chang, Y.-X. Cheng, Z. Zuo, P. Li, Discrimination of neutrons and gamma-rays in plastic scintillator based on spiking cortical model, Nuclear Engineering and Technology 55 (9) (2023) 3359–3366. doi:10.1016/ j.net.2023.04.032

    2023

  4. [12]

    Fabian, G

    X. Fabian, G. Baulieu, L. Ducroux, O. St´ ezowski, A. Boujrad, E. Cl´ ement, S. Coudert, G. de France, N. Erduran, S. Ert¨ urk, V. Gonz´ alez, G. Jaworski, J. Nyberg, D. Ralet, E. Sanchis, R. Wadsworth, Artificial neural networks for neutron/ γ discrimination in the neutron de...

    2021

  5. [13]

    S. Yoon, C. Lee, H. Seo, H.-D. Kim, Improved fast neutron detection using CNN-based pulse shape discrimination, Nuclear Engineering and Technology 55 (11) (2023) 3925–3934. doi:10.1016/j.net.2023.07.007

    2023 doi

  6. [14]

    K. Zhao, C. Feng, S. Wang, Z. Shen, K. Zhang, S. Liu, n/ γ discrimination for CLYC detector using a one-dimensional Convolutional Neural Network, Journal of Instrumentation 18 (01) (2023) P01021.doi:10.1088/1748-0221/18/01/P01021

    2023 doi

  7. [15]

    Doucet, T

    E. Doucet, T. Brown, P. Chowdhury, C. Lister, C. Morse, P. Bender, A. Rogers, Machine learning n/ γ discrimination in CLYC scintillators, Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment 954 (2020) 1...

    2020 doi

  8. [16]

    Hachem, Y

    A. Hachem, Y. Moline, G. Corre, B. Ouni, M. Trocme, A. Elayeb, F. Carrel, Labeling strategy to improve neutron/gamma discrimination with organic scintillator, Nuclear Engineering and Technology 55 (11) (2023) 4057–4065. doi:10.1016/j. net.2023.07.024

    2023 doi

  9. [17]

    Zhang, Z

    S. Zhang, Z. Wei, P. Zhang, Q. Zhao, M. Li, X. Bai, K. Wu, Y. Nie, Y. Ding, J. Wang, Y. Zhang, X. Su, Z. Yao, Neutron- gamma discrimination with broaden the lower limit of energy threshold using BP neural network, Applied Radiation and Isotopes 205 (2024) 111179. doi:10.1016/j...

    2024

  10. [18]

    V. H. Hai, N. M. Dang, N. T. T. Phuc, H. T. K. Trang, T. T. H. Loan, P. L. H. Sang, M. Nomachi, Enhancing Neu- tron/Gamma Discrimination in the Low-Energy Region for EJ-276 Plastic Scintillation Detector Using Machine Learning, IEEE Transactions on Nuclear Science (2024) 1–1 d...

    2024

  11. [19]

    Y. Pan, P. Gong, Z. Hu, Z. Wang, D. Liang, C. Zhou, X. Zhu, X. Tang, Pulse pile-up recognition using multi-module DenseNet in neutron-gamma discrimination, Nuclear Engineering and Technology (2024) 103329doi:10.1016/j.net.2024. 11.031

    2024 doi

  12. [20]

    J. Han, J. Zhu, Z. Wang, G. Qu, X. Liu, W. Lin, Z. Xu, Y. Huang, M. Yan, X. Zhang, L. Chen, Pulse characteristics of CLYC and piled-up neutron–gamma discrimination using a convolutional neural network, Nuclear Instruments and Methods in Physics Research Section A: Accelerators...

    2022

  13. [21]

    S. Peng, Z. Hua, Q. Wu, J. Han, S. Qian, Z. Wang, Q. Wei, L. Qin, L. Ma, M. Yan, R. Song, Piled-up neutron-gamma discrimination system for CLLB using convolutional neural network, Journal of Instrumentation 17 (08) (2022) T08001. doi:10.1088/1748-0221/17/08/T08001

    2022 doi

  14. [22]

    I. R. Morales, M. L. Crespo, M. Bogovac, A. Cicuttin, K. Kanaki, S. Carrato, Gamma/neutron classification with SiPM CLYC detectors using frequency-domain analysis for embedded real-time applications, Nuclear Engineering and Technology 56 (2) (2024) 745–752. doi:10.1016/j.net.2...

    2024 doi

  15. [23]

    Argueta, I

    M. Argueta, I. Ren´ e, Gamma and neutron tagged dataset from CLYC SiPM detector (2023).doi:10.5281/zenodo.8037239

    2023 doi

  16. [24]

    V. Nair, G. E. Hinton, Rectified linear units improve restricted boltzmann machines, in: J. F¨ urnkranz, T. Joachims (Eds.), Proceedings of the 27th International Conference on Machine Learning (ICML-10), June 21-24, 2010, Haifa, Israel, Omnipress, 2010, pp. 807–814

    2010

  17. [25]

    Alharbi, Distance metrics for digital pulse-shape discrimination of scintillator detectors, Radiation Physics and Chem- istry 156 (2019) 205–209

    T. Alharbi, Distance metrics for digital pulse-shape discrimination of scintillator detectors, Radiation Physics and Chem- istry 156 (2019) 205–209. doi:10.1016/j.radphyschem.2018.11.014

    2019 doi

  18. [26]

    7B30_SIP-E3-CLYC-X.pdf , accessed: 2025-04-10

    Thermal NEUTRON detector V12.7B30/SIP-E3-CLYC-X, https://scionix.nl/wp-content/uploads/2017/03/V12. 7B30_SIP-E3-CLYC-X.pdf , accessed: 2025-04-10

    2017

  19. [27]

    D. P. Kingma, J. Ba, Adam: A method for stochastic optimization, in: Y. Bengio, Y. LeCun (Eds.), 3rd International Conference on Learning Representations, ICLR 2015, San Diego, CA, USA, May 7-9, 2015, Conference Track Proceedings, 2015

    2015

  20. [28]

    Paszke, S

    A. Paszke, S. Gross, F. Massa, A. Lerer, J. Bradbury, G. Chanan, T. Killeen, Z. Lin, N. Gimelshein, L. Antiga, A. Des- maison, A. K¨ opf, E. Z. Yang, Z. DeVito, M. Raison, A. Tejani, S. Chilamkurthy, B. Steiner, L. Fang, J. Bai, S. Chintala, Pytorch: An imperative style, high-...

    2019

  21. [29]

    Cesium Yttrium Chloride (CLYC:Ce) for Dual Neutron & Gamma Radiation Detection, https://www.symmic.net/docs/ CLYC.pdf, accessed: 2025-04-10

    2025

  22. [30]

    C-Series SiPM Sensors, https://www.onsemi.cn/download/data-sheet/pdf/microc-series-d.pdf , accessed: 2025-04-10

    2025

  23. [31]

    P. Li, H. Liu, Dataset for neutron and gamma-ray pulse shape discrimination: radiation pulse signals and discrimination methodologies (2024). doi:10.5281/zenodo.10947029

    2024 doi

  24. [32]

    Maedgen, B

    P. Maedgen, B. Wellons, S. Prasad, J. Tao, Improving Pulse Shape Discrimination in Organic Scintillation Detectors by Understanding Underlying Data Structure (dataset), https://github.com/NeutronNeutrinoSensing/PSDwithML, accessed: 2025-04-10. 16

    2025

Pith tools

Reviewed August 8, 2026 · model on record in the stance chip above.