pith. sign in

arxiv: 2412.20091 · v6 · submitted 2024-12-28 · 🌌 astro-ph.HE

Gamma-Ray Burst Light Curve Reconstruction: A Comparative Machine and Deep Learning Analysis

A. Manchanda , A. Kaushal , M. G. Dainotti , A. Deepu , S. Naqi , J. Felix , N. Indoriya , S. P. Magesh
show 9 more authors
H. Gupta K. Gupta A. Madhan D. H. Hartmann A. Pollo M. Bogdan J. X. Prochaska N. Fraija D. Debnath
This is my paper

Pith reviewed 2026-05-23 07:15 UTC · model grok-4.3

classification 🌌 astro-ph.HE
keywords Gamma-ray burstsLight curve reconstructionMachine learningDainotti relationPlateau phaseMLPAttention U-NetCosmological tools
0
0 comments X

The pith

The multi-layer perceptron reduces uncertainties in GRB plateau parameters by 37 to 41 percent while achieving the lowest mean squared error among nine tested models.

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

Gamma-ray bursts can serve as cosmological probes if their parameter correlations have low dispersion. Gaps in their light curves, especially around the plateau phase, increase uncertainties in key parameters like the plateau end time and luminosity. This paper tests nine machine learning and statistical models to reconstruct those gaps on a sample of 521 bursts. The multi-layer perceptron achieves the lowest mean squared error and cuts uncertainties in the plateau parameters by 37 to 41 percent relative to the standard Willingale fit. Attention U-Net reduces uncertainties the most but has higher error, so the MLP is presented as the more reliable choice for tightening the Dainotti relation.

Core claim

Applying nine models including multi-layer perceptrons, recurrent networks, generative adversarial networks, and attention-based networks to reconstruct gaps in 521 gamma-ray burst light curves shows that the multi-layer perceptron attains the lowest 5-fold cross-validation mean squared error of 0.0275. It reduces uncertainties in the logarithm of the plateau end time by 37.2 percent, in the logarithm of the plateau flux by 38.0 percent, and in the post-plateau slope by 41.2 percent relative to the Willingale model. The attention U-Net produces the greatest uncertainty reductions of 37.9, 38.5, and 41.4 percent respectively but records a higher mean squared error of 0.134, rendering the MLP

What carries the argument

The multi-layer perceptron (MLP) applied to light curve gap reconstruction to minimize errors in Dainotti relation parameters.

If this is right

  • Reduced parameter uncertainties tighten the dispersion in the Dainotti relation between plateau end time and luminosity.
  • More precise parameters enhance the utility of gamma-ray bursts as standard candles for cosmological distance measurements.
  • The reconstruction approach supports the development of machine learning tools for estimating gamma-ray burst redshifts.
  • Improved fits to light curves facilitate tests of theoretical models for the plateau emission phase.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same reconstruction pipeline could be adapted to other astrophysical transients that exhibit similar plateau or gap features in their light curves.
  • Systematic comparison against fully observed bursts from future wide-field monitors would quantify any residual biases introduced by the models.
  • Combining the low-error MLP with the low-uncertainty attention U-Net in an ensemble might yield further gains in both accuracy and precision.
  • Application to real-time burst alerts could accelerate redshift predictions and follow-up observations.

Load-bearing premise

The reconstructed light curve segments accurately recover the true underlying emission physics without introducing systematic biases that propagate into the Dainotti relation parameters or cosmological inferences.

What would settle it

A direct test on a set of simulated gamma-ray burst light curves with known true plateau parameters would reveal whether the reported uncertainty reductions correspond to actual improvements in parameter recovery accuracy.

Figures

Figures reproduced from arXiv: 2412.20091 by A. Deepu, A. Kaushal, A. Madhan, A. Manchanda, A. Pollo, D. Debnath, D. H. Hartmann, H. Gupta, J. Felix, J. X. Prochaska, K. Gupta, M. Bogdan, M. G. Dainotti, N. Fraija, N. Indoriya, S. Naqi, S. P. Magesh.

Figure 1
Figure 1. Figure 1: Row 1 and Row 2 describe the GRB LCs divided into four classes depending on the afterglow feature: i) Good GRBs (Row 1, left); ii) GRB LCs with a break towards the end (Row 1, right); iii) Flares or Bumps in the afterglow (Row 2, left); iv) Flares or Bumps with a Double Break towards the end of the LC (Row 3, right). Row 3’s left plot shows the LC of GRB050822, starting from the plateau emission, and the b… view at source ↗
Figure 2
Figure 2. Figure 2: Comparison of LCs before and after hyperparameter tuning. combination that results in the best performance according to a chosen metric. The n restarts optimizer is set to 20 and the alpha hyperparameter is set as 0.1. After tuning the hyperparameters, we fit the GP Regressor to the data. The optimized model improved the recon￾struction of the LCs, as shown in Fig. 2b. It might appear that Fig. 2a offers a… view at source ↗
Figure 3
Figure 3. Figure 3: Architecture of nine models shown from top left to bottom right in each row: (a) MLP, (b) Attention U-Net, (c) Bi￾LSTM, (d) Bi-MAMBA, (e) GP-RF, (f) CGAN, (g) KAN, (h) Fourier Transform, (i) SARIMAX-based Kalman . Architectures of Bi-Mamba, MLP, KAN, and Attention U-Net are actual representations of the models utilized [PITH_FULL_IMAGE:figures/full_fig_p013_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Reconstruction of LCs using the W07 function. Row 1 shows the reconstructed LCs at 10% noise level and Row 2 shows the reconstructed LCs at 20% noise level for all four classes: i) Good GRBs (Column 1); ii) a break towards the end of the GRB LC (Column 2); iii) flares or bumps in the afterglow (Column 3); iv) flares or bumps with a double break towards the end of the LC (Column 4) [PITH_FULL_IMAGE:figures… view at source ↗
Figure 5
Figure 5. Figure 5: Reconstruction of LCs for all four varieties of GRBs are shown in a grid with four types of GRBs (left to right): i) Good GRBs (Column 1); ii) a GRB LC with a break towards the end (Column 2); iii) flares or bumps in the afterglow (Column 3); iv) flares or bumps with a double break towards the end of the LC (Column 4) and the models (top to bottom): i) GP (Row 1); ii) Bi-Mamba Model (Row 2); iii) MLP model… view at source ↗
Figure 6
Figure 6. Figure 6: Reconstruction of LCs for all four varieties of GRBs are shown in a grid with four types of GRBs (left to right): i) Good GRBs (Column 1); ii) a GRB LC with a break towards the end (Column 2); iii) flares or bumps in the afterglow (Column 3); iv) flares or bumps with a double break towards the end of the LC (Column 4) and the models (top to bottom): i) Bi-LSTM Model(Row 1); ii) SARIMA-based Kalman Model (R… view at source ↗
Figure 7
Figure 7. Figure 7: Distribution plot of all three W07 parameters in a grid with parameters (left to right): i) log Fa (Column 1) ii) log Ta (Column 2) iii) α (Column 3) and the models (top to bottom): i) Willingale (10% noise) (Row 1); ii) W07 (20% noise) (Row 2); iii) GP Model(Row 3); iv) Bi-Mamba Model (Row 4); v) MLP Model (Row 5); vi) Fourier (Row 6) [PITH_FULL_IMAGE:figures/full_fig_p025_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Distribution plot of all three W07 parameters in a grid with parameters (left to right): i) log Fa (Column 1) ii) log Ta (Column 2) iii) α (Column 3) and the models (top to bottom): i) GP-RF (Row 1); ii) Bi-LSTM (Row 2); iii) CGAN (Row 3); iv) SARIMAX-based Kalman (Row 4); v) KAN (Row 5); vi) Attention U-Net (Row 6) [PITH_FULL_IMAGE:figures/full_fig_p026_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: Reconstruction of LCs using Bi-Mamba trained on GRB050713A. Left: Reconstructed LC of GRB241026A; Right: Reconstructed LC of GRB241127A. 6.2. TimeAutoDiff TimeAutoDiff (Suh et al. 2024) is a model designed to synthesize time-series tabular data, addressing the challenges of temporal dependencies and heterogeneous features. It combines a Variational Autoencoder (VAE) (Kingma et al. 2019) with a Denoising Di… view at source ↗
Figure 10
Figure 10. Figure 10: Reconstructed GRB LCs using i) Random Forest model (left); ii) Incorporation of both Random Forest and GP Regression Models (right). 6.6. The Hybrid Model: Random Forest with Gaussian Process To counter the step-like nature of Random Forest predictions in LCR, we develop a hybrid model by combining Random Forest with a GP. Here, the Random Forest acts as the base model, trained on the data to capture init… view at source ↗
read the original abstract

Gamma-Ray Bursts (GRBs), observed at high-z, are probes of the evolution of the Universe and can be used as cosmological tools. Thus, we need correlations with small dispersion among key parameters. To reduce such a dispersion, we mitigate gaps in light curves (LCs), including the plateau region, key to building the two-dimensional Dainotti relation between the end time of plateau emission (Ta) and its luminosity (La). We reconstruct LCs using nine models: Multi-Layer Perceptron (MLP), Bi-Mamba, Fourier Transform, Gaussian Process-Random Forest Hybrid (GP-RF), Bidirectional Long Short-Term Memory (Bi-LSTM), Conditional GAN (CGAN), SARIMAX-based Kalman filter, Kolmogorov-Arnold Networks (KANs), and Attention U-Net. These methods are compared to the Willingale model (W07) over a sample of 521 GRBs. MLP and Attention U-Net outperform other methods, with MLP reducing the plateau parameter uncertainties by 37.2% for log Ta, 38.0% for log Fa, and 41.2% for alpha (the post-plateau slope in the W07 model), achieving the lowest 5-fold cross-validation (CV) mean squared error (MSE) of 0.0275. Attention U-Net achieved the lowest uncertainty of parameters, a 37.9% reduction in log Ta, a 38.5% reduction in log Fa and a 41.4% reduction in alpha, but with a higher MSE of 0.134. Although Attention U-Net achieves the largest uncertainty reduction, the MLP attains the lowest test MSE while maintaining comparable uncertainty performance, making it the more reliable model. The other methods yield MSE values ranging from 0.0339 to 0.174. These improvements in parameter precision are needed to use GRBs as standard candles, investigate theoretical models, and predict GRB redshifts through machine learning.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

3 major / 2 minor

Summary. The paper compares nine ML/DL models (MLP, Bi-Mamba, Fourier Transform, GP-RF, Bi-LSTM, CGAN, SARIMAX-Kalman, KANs, Attention U-Net) against the Willingale 2007 (W07) model for reconstructing GRB light curves on 521 events, with emphasis on recovering plateau parameters (Ta, Fa, alpha) to tighten the Dainotti relation. MLP is reported to achieve the lowest 5-fold CV MSE (0.0275) and uncertainty reductions of 37.2–41.2% relative to W07; Attention U-Net yields the largest uncertainty reductions but higher MSE (0.134). The work positions these reconstructions as enabling better cosmological use of GRBs.

Significance. If the reconstructed parameters prove unbiased, the reported precision gains could meaningfully reduce scatter in the Dainotti relation and support GRB cosmology. The explicit 5-fold CV comparison to an external baseline (W07) and the large sample size are strengths; however, the absence of ground-truth simulations means the significance hinges on an untested assumption that lower MSE and smaller reported uncertainties correspond to faithful recovery of the underlying emission parameters.

major comments (3)
  1. [§4] §4 (Results) and abstract: the central performance claims rest on 5-fold CV MSE computed on held-out observed flux points, yet no validation is presented against simulated light curves with known true Ta, Fa, and alpha values. This metric therefore cannot distinguish between reduced scatter around the observed points and systematic shifts in the fitted plateau parameters that would propagate into the Dainotti relation slope or intrinsic dispersion.
  2. [§4.2] §4.2 and Table 2 (parameter uncertainties): uncertainty reductions (e.g., 37.2% for log Ta under MLP) are reported relative to W07, but the manuscript contains no explicit check that the new parameter sets preserve the slope or reduce the scatter of the log La–log Ta correlation itself—the stated scientific motivation. Without this test the claimed cosmological utility remains unverified.
  3. [§3] Methods (§3) and abstract: post-hoc designation of MLP as the preferred model (lowest MSE) while Attention U-Net shows larger uncertainty reductions but higher MSE (0.134) lacks a pre-specified decision rule or joint figure-of-merit; this selection risks inflating apparent performance on the reported metrics.
minor comments (2)
  1. [§2] Notation for the post-plateau slope (alpha) should be defined once in §2 and used consistently; its relation to the W07 functional form is not restated in the results tables.
  2. [Figures] Figure captions for the example light-curve reconstructions should state the number of GRBs shown and whether the plotted curves are mean predictions or individual realizations.

Simulated Author's Rebuttal

3 responses · 0 unresolved

Thank you for the constructive feedback. We respond to each major comment below, proposing revisions to address the concerns raised.

read point-by-point responses

  1. Referee: [§4] §4 (Results) and abstract: the central performance claims rest on 5-fold CV MSE computed on held-out observed flux points, yet no validation is presented against simulated light curves with known true Ta, Fa, and alpha values. This metric therefore cannot distinguish between reduced scatter around the observed points and systematic shifts in the fitted plateau parameters that would propagate into the Dainotti relation slope or intrinsic dispersion.

    Authors: We concur that simulated light curves with known ground-truth parameters would offer direct validation of unbiased recovery. Since our analysis uses real GRB observations where true values are unavailable, the 5-fold CV on held-out points assesses predictive accuracy on observed data. We will revise the manuscript to explicitly discuss this limitation in a new subsection and emphasize that the reported gains are conditional on the assumption of no systematic bias. Future extensions will incorporate simulations. revision: partial

  2. Referee: [§4.2] §4.2 and Table 2 (parameter uncertainties): uncertainty reductions (e.g., 37.2% for log Ta under MLP) are reported relative to W07, but the manuscript contains no explicit check that the new parameter sets preserve the slope or reduce the scatter of the log La–log Ta correlation itself—the stated scientific motivation. Without this test the claimed cosmological utility remains unverified.

    Authors: While the focus was on uncertainty reduction in individual parameters, we recognize the importance of verifying the impact on the Dainotti relation. In the revised manuscript, we will add an analysis comparing the slope and intrinsic scatter of the log La–log Ta relation using parameters from W07 and from the MLP reconstructions. revision: yes

  3. Referee: [§3] Methods (§3) and abstract: post-hoc designation of MLP as the preferred model (lowest MSE) while Attention U-Net shows larger uncertainty reductions but higher MSE (0.134) lacks a pre-specified decision rule or joint figure-of-merit; this selection risks inflating apparent performance on the reported metrics.

    Authors: The selection of MLP was based on achieving the lowest MSE, with Attention U-Net noted for its uncertainty performance. To address the lack of a pre-specified rule, we will update the Methods section to define the model selection criteria upfront, using MSE as the primary metric and uncertainty reduction as secondary. A combined figure-of-merit will be introduced in the results for transparency. revision: yes

Circularity Check

0 steps flagged

No significant circularity in reconstruction metrics or parameter comparisons

full rationale

The paper's claims rest on empirical 5-fold CV MSE computed on held-out observed flux points in the light curves and on uncertainty reductions measured relative to the independent external W07 baseline model across 521 GRBs. These quantities are not defined in terms of the target Dainotti parameters themselves, nor do any steps reduce by construction to fitted inputs or self-citation chains. The derivation is self-contained against external benchmarks with no load-bearing self-definitional, fitted-prediction, or ansatz-smuggling patterns.

Axiom & Free-Parameter Ledger

1 free parameters · 1 axioms · 0 invented entities

Abstract-only review limits visibility into exact modeling choices; ML models inherently involve many fitted parameters and the baseline W07 model is treated as given.

free parameters (1)
  • ML model hyperparameters
    Each of the nine models requires tuning of architecture and training parameters to the GRB dataset.
axioms (1)
  • domain assumption GRB light curves contain a well-defined plateau phase whose parameters follow the Willingale 2007 functional form
    Used as the reference model against which reconstructions are compared.

pith-pipeline@v0.9.0 · 5980 in / 1292 out tokens · 22484 ms · 2026-05-23T07:15:57.244141+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 2 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. NucleiML: A machine learning framework of ground-state properties of finite nuclei for accelerated Bayesian exploration

    nucl-th 2025-04 unverdicted novelty 5.0

    NucleiML is a machine learning surrogate for relativistic mean-field calculations of finite nuclei properties that accelerates Bayesian inference of nuclear equation of state parameters by roughly 1000 times.

  2. Gamma-Ray Bursts as an Independent High-Redshift Probe of Dark Energy

    astro-ph.CO 2026-03 unverdicted novelty 4.0

    Forecasts show that ~66 optical GRBs can achieve σ_w ≈ 0.47 in wCDM using Dainotti relations, matching Planck precision and enabling independent high-redshift tests of dark energy.

Reference graph

Works this paper leans on

108 extracted references · 108 canonical work pages · cited by 2 Pith papers · 15 internal anchors

  1. [1]

    Deep Learning using Rectified Linear Units (ReLU)

    Agarap, A. 2018, arXiv preprint arXiv:1803.08375

    work page internal anchor Pith review Pith/arXiv arXiv 2018

  2. [2]

    2019, in Proceedings of the 25th ACM SIGKDD international conference on knowledge discovery & data mining, 2623–2631

    Akiba, T., Sano, S., Yanase, T., Ohta, T., & Koyama, M. 2019, in Proceedings of the 25th ACM SIGKDD international conference on knowledge discovery & data mining, 2623–2631

    work page 2019

  3. [3]

    R., & Csala, D

    Alharbi, F. R., & Csala, D. 2022, Inventions, 7, doi: 10.3390/inventions7040094

    work page doi:10.3390/inventions7040094 2022

  4. [4]

    2018, Advances in Space Research, 62, 191

    Amati, L., O’Brien, P., G¨ otz, D., et al. 2018, Advances in Space Research, 62, 191

    work page 2018

  5. [5]

    Aravkin, A., Burke, J., & Pillonetto, G. 2013

    work page 2013

  6. [6]

    2022, International Journal of Modern Physics D, 31, 2230008

    Atteia, J.-L., Cordier, B., & Wei, J. 2022, International Journal of Modern Physics D, 31, 2230008

    work page 2022

  7. [7]

    G., & Capozziello, S

    Bargiacchi, G., Dainotti, M. G., & Capozziello, S. 2025, New Astronomy Reviews, 100, 101712

    work page 2025

  8. [8]

    D., Barbier, L

    Barthelmy, S. D., Barbier, L. M., Cummings, J. R., et al. 2005, Space Science Reviews, 120, 143, doi: 10.1007/s11214-005-5096-3

    work page internal anchor Pith review doi:10.1007/s11214-005-5096-3 2005

  9. [9]

    2010, The Astrophysical Journal Letters, 719, L10, doi: 10.1088/2041-8205/719/1/L10

    Beskin, G., Karpov, S., Bondar, S., et al. 2010, The Astrophysical Journal Letters, 719, L10, doi: 10.1088/2041-8205/719/1/L10

    work page doi:10.1088/2041-8205/719/1/l10 2010

  10. [10]

    2014, Astronomy &; Astrophysics

    Betoule, M., Kessler, R., Guy, J., et al. 2014, Astronomy &; Astrophysics

    work page 2014

  11. [11]

    H., Bloom, J

    Blake, C. H., Bloom, J. S., Starr, D. L., et al. 2005, Nature, 435, 181, doi: 10.1038/nature03520

    work page doi:10.1038/nature03520 2005

  12. [12]

    2001, Machine learning, 45, 5

    Breiman, L. 2001, Machine learning, 45, 5

    work page 2001

  13. [13]

    A., & Stone, C

    Breiman, L., Friedman, J., Olshen, R. A., & Stone, C. J. 2017, Classification and regression trees (Routledge)

    work page 2017

  14. [14]

    2023, ACM Computing Surveys, 55, 1

    Brophy, E., Wang, Z., She, Q., & Ward, T. 2023, ACM Computing Surveys, 55, 1

    work page 2023

  15. [15]

    N., Hill, J

    Burrows, D. N., Hill, J. E., Nousek, J. A., et al. 2005, Space Science Reviews, 120, 165, doi: 10.1007/s11214-005-5097-2

    work page internal anchor Pith review doi:10.1007/s11214-005-5097-2 2005

  16. [16]

    2022a, MNRAS, 512, 439, doi: 10.1093/mnras/stac517

    Cao, S., Dainotti, M., & Ratra, B. 2022a, MNRAS, 512, 439, doi: 10.1093/mnras/stac517

    work page doi:10.1093/mnras/stac517

  17. [17]

    2022b, MNRAS, 510, 2928, doi: 10.1093/mnras/stab3559

    Cao, S., Khadka, N., & Ratra, B. 2022b, MNRAS, 510, 2928, doi: 10.1093/mnras/stab3559

    work page doi:10.1093/mnras/stab3559

  18. [18]

    R., 2009, @doi [ ] 10.1111/j.1365-2966.2009.15167.x , http://adsabs.harvard.edu/abs/2009MNRAS.398..607V 398, 607

    Cardone, V. F., Capozziello, S., & Dainotti, M. G. 2009, Monthly Notices of the Royal Astronomical Society, 400, 775, doi: 10.1111/j.1365-2966.2009.15456.x

    work page doi:10.1111/j.1365-2966.2009.15456.x 2009

  19. [19]

    J., Reeves J

    Willingale, R. 2010, Monthly Notices of the Royal Astronomical Society, 408, 1181, doi: 10.1111/j.1365-2966.2010.17197.x

    work page doi:10.1111/j.1365-2966.2010.17197.x 2010

  20. [20]

    Long Short-Term Memory-Networks for Machine Reading

    Cheng, J., Dong, L., & Lapata, M. 2016, arXiv preprint arXiv:1601.06733

    work page internal anchor Pith review Pith/arXiv arXiv 2016

  21. [21]

    1997, Nature, 387, 783, doi: 10.1038/42885

    Costa, E., Frontera, F., Heise, J., et al. 1997, Nature, 387, 783, doi: 10.1038/42885

    work page doi:10.1038/42885 1997

  22. [22]

    J., Fox, D

    Cucchiara, A., Levan, A. J., Fox, D. B., et al. 2011, The Astrophysical Journal, 736, 7, doi: 10.1088/0004-637X/736/1/7

    work page doi:10.1088/0004-637x/736/1/7 2011

  23. [23]

    2024a, arXiv preprint arXiv:2411.10736 35

    Dainotti, M., Bhardwaj, S., Bissaldi, E., et al. 2024a, arXiv preprint arXiv:2411.10736 35

    work page arXiv

  24. [24]

    2024b, arXiv preprint arXiv:2401.03589

    Dainotti, M., Taira, E., Wang, E., et al. 2024b, arXiv preprint arXiv:2401.03589

    work page arXiv

  25. [25]

    Tavecchio and G

    Dainotti, M. G., Cardone, V. F., & Capozziello, S. 2008, Monthly Notices of the Royal Astronomical Society, 391, L79, doi: 10.1111/j.1745-3933.2008.00560.x

    work page doi:10.1111/j.1745-3933.2008.00560.x 2008

  26. [26]

    2015, The Astrophysical Journal, 800, 31, doi: 10.1088/0004-637X/800/1/31

    Capozziello, S. 2015, The Astrophysical Journal, 800, 31, doi: 10.1088/0004-637X/800/1/31

    work page doi:10.1088/0004-637x/800/1/31 2015

  27. [27]

    2011, The Astrophysical Journal, 730, 135, doi: 10.1088/0004-637X/730/2/135

    Ostrowski, M., & Willingale, R. 2011, The Astrophysical Journal, 730, 135, doi: 10.1088/0004-637X/730/2/135

    work page doi:10.1088/0004-637x/730/2/135 2011

  28. [28]

    G., Lenart, A

    Dainotti, M. G., Lenart, A. L., Fraija, N., et al. 2021, Publications of the Astronomical Society of Japan, 73, 970

    work page 2021

  29. [29]

    G., Lenart, A

    Dainotti, M. G., Lenart, A. L., Sarracino, G., et al. 2020a, The Astrophysical Journal, 904, 97, doi: 10.3847/1538-4357/abbe8a

    work page doi:10.3847/1538-4357/abbe8a

  30. [30]

    2017, Astronomy and Astrophysics, 600, A98, doi: 10.1051/0004-6361/201628384

    Pian, E. 2017, Astronomy and Astrophysics, 600, A98, doi: 10.1051/0004-6361/201628384

    work page doi:10.1051/0004-6361/201628384 2017

  31. [31]

    G., Nielson, V., Sarracino, G., et al

    Dainotti, M. G., Nielson, V., Sarracino, G., et al. 2022g, MNRAS, 514, 1828, doi: 10.1093/mnras/stac1141

    work page doi:10.1093/mnras/stac1141

  32. [32]

    G., Petrosian, V., Singal, J., & Ostrowski, M

    Dainotti, M. G., Petrosian, V., Singal, J., & Ostrowski, M. 2013, The Astrophysical Journal, 774, 157, doi: 10.1088/0004-637X/774/2/157

    work page doi:10.1088/0004-637x/774/2/157 2013

  33. [33]

    2016, The Astrophysical Journal Letters, 825, L20, doi: 10.3847/2041-8205/825/2/L20

    Ostrowski, M. 2016, The Astrophysical Journal Letters, 825, L20, doi: 10.3847/2041-8205/825/2/L20

    work page doi:10.3847/2041-8205/825/2/l20 2016

  34. [34]

    G., Postnikov, S., Hernandez, X., & Ostrowski, M

    Dainotti, M. G., Postnikov, S., Hernandez, X., & Ostrowski, M. 2016, The Astrophysical Journal Letters, 825, L20

    work page 2016

  35. [35]

    G., Singal, J., Ostrowski, M., et al

    Dainotti, M. G., Singal, J., Ostrowski, M., et al. 2013, The Astrophysical Journal, 774, 157

    work page 2013

  36. [36]

    2010, The Astrophysical Journal Letters, 722, L215, doi: 10.1088/2041-8205/722/2/L215

    Cardone, V., & Ostrowski, M. 2010, The Astrophysical Journal Letters, 722, L215, doi: 10.1088/2041-8205/722/2/L215

    work page doi:10.1088/2041-8205/722/2/l215 2010

  37. [37]

    G., Livermore, S., Kann, D

    Dainotti, M. G., Livermore, S., Kann, D. A., et al. 2020b, The Astrophysical Journal Letters, 905, L26, doi: 10.3847/2041-8213/abcda9

    work page doi:10.3847/2041-8213/abcda9 2041

  38. [38]

    G., Young, S., Li, L., et al

    Dainotti, M. G., Young, S., Li, L., et al. 2022, ApJS, 261, 25, doi: 10.3847/1538-4365/ac7c64

    work page doi:10.3847/1538-4365/ac7c64 2022

  39. [39]

    2023, Astronomy & Astrophysics, 677, A16 Dereli-B´ egu´ e, H., Pe’er, A., B´ egu´ e, D., & Ryde, F

    Demianenko, M., Malanchev, K., Samorodova, E., et al. 2023, Astronomy & Astrophysics, 677, A16 Dereli-B´ egu´ e, H., Pe’er, A., B´ egu´ e, D., & Ryde, F. 2024, arXiv preprint arXiv:2412.11533

    work page arXiv 2023

  40. [40]

    2009, Monthly Notices of the Royal Astronomical Society, 397, 1177

    Evans, P., Beardmore, A., Page, K., et al. 2009, Monthly Notices of the Royal Astronomical Society, 397, 1177

    work page 2009

  41. [41]

    Time Series Simulation by Conditional Generative Adversarial Net

    Fu, R., Chen, J., Zeng, S., Zhuang, Y., & Sudjianto, A. 2019, arXiv preprint arXiv:1904.11419

    work page internal anchor Pith review Pith/arXiv arXiv 2019

  42. [42]

    Gehrels, N., Ramirez-Ruiz, E., & Fox, D. B. 2009, Annual Review of Astronomy and Astrophysics, 47, 567, doi: 10.1146/annurev.astro.46.060407.145147

    work page doi:10.1146/annurev.astro.46.060407.145147 2009

  43. [43]

    Gehrels, G

    Gehrels, N., Chincarini, G., Giommi, P., et al. 2004, The Astrophysical Journal, 611, 1005, doi: 10.1086/422091

    work page doi:10.1086/422091 2004

  44. [44]

    J., Pouget-Abadie, J., Mirza, M., et al

    Goodfellow, I. J., Pouget-Abadie, J., Mirza, M., et al. 2014, Advances in neural information processing systems, 27

    work page 2014

  45. [45]

    and Shibahashi , H

    Gorbovskoy, E. S., Lipunova, G. V., Lipunov, V. M., et al. 2012, Monthly Notices of the Royal Astronomical Society, 421, 1874, doi: 10.1111/j.1365-2966.2012.20195.x

    work page doi:10.1111/j.1365-2966.2012.20195.x 2012

  46. [46]

    2005, in International conference on artificial neural networks, Springer, 799–804

    Graves, A., Fern´ andez, S., & Schmidhuber, J. 2005, in International conference on artificial neural networks, Springer, 799–804

    work page 2005

  47. [47]

    Mamba: Linear-Time Sequence Modeling with Selective State Spaces

    Gu, A., & Dao, T. 2024, Mamba: Linear-Time Sequence Modeling with Selective State Spaces. https://arxiv.org/abs/2312.00752

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  48. [48]

    Delving Deep into Rectifiers:

    He, K., Zhang, X., Ren, S., & Sun, J. 2015, in 2015 IEEE International Conference on Computer Vision (ICCV), 1026–1034, doi: 10.1109/ICCV.2015.123

    work page doi:10.1109/iccv.2015.123 2015

  49. [49]

    2018, Genetic programming and evolvable machines, 19, 305

    Heaton, J. 2018, Genetic programming and evolvable machines, 19, 305

    work page 2018

  50. [50]

    2020, Advances in neural information processing systems, 33, 6840

    Ho, J., Jain, A., & Abbeel, P. 2020, Advances in neural information processing systems, 33, 6840

    work page 2020

  51. [51]

    1998, Computers & Industrial Engineering, 35, 213, doi: https://doi.org/10.1016/S0360-8352(98)00066-7

    Ho, S., & Xie, M. 1998, Computers & Industrial Engineering, 35, 213, doi: https://doi.org/10.1016/S0360-8352(98)00066-7

    work page doi:10.1016/s0360-8352(98)00066-7 1998

  52. [52]

    Long short -term memory,

    Hochreiter, S., & Schmidhuber, J. 1997, Neural computation, 9, 1735, doi: 10.1162/neco.1997.9.8.1735

    work page doi:10.1162/neco.1997.9.8.1735 1997

  53. [53]

    2019, IEEE Communications Magazine, 57, 114

    Hua, Y., Zhao, Z., Li, R., et al. 2019, IEEE Communications Magazine, 57, 114

    work page 2019

  54. [54]

    2024, arXiv preprint arXiv:2403.08029 36

    Huber, S., & Suyu, S. 2024, arXiv preprint arXiv:2403.08029 36

    work page arXiv 2024

  55. [55]

    Adam: A Method for Stochastic Optimization

    Kingma, D. P., & Ba, J. 2014, Adam: A Method for Stochastic Optimization. https://arxiv.org/abs/1412.6980

    work page internal anchor Pith review Pith/arXiv arXiv 2014

  56. [56]

    P., Welling, M., et al

    Kingma, D. P., Welling, M., et al. 2019, Foundations and Trends®in Machine Learning, 12, 307

    work page 2019

  57. [57]

    1995, in Proceedings of the 14th International Joint Conference on Artificial Intelligence (IJCAI), Vol

    Kohavi, R. 1995, in Proceedings of the 14th International Joint Conference on Artificial Intelligence (IJCAI), Vol. 14

    work page 1995

  58. [58]

    Kumar, P., & Duran, R. B. 2010, Monthly Notices of the Royal Astronomical Society, 409, 226

    work page 2010

  59. [59]

    2015, PhR, 561, 1, doi: 10.1016/j.physrep.2014.09.008

    Kumar, P., & Zhang, B. 2015, PhR, 561, 1, doi: 10.1016/j.physrep.2014.09.008

    work page internal anchor Pith review doi:10.1016/j.physrep.2014.09.008 2015

  60. [60]

    J., et al

    Levine, D., Dainotti, M., Zvonarek, K. J., et al. 2022, The Astrophysical Journal, 925, 15, doi: 10.3847/1538-4357/ac4221

    work page doi:10.3847/1538-4357/ac4221 2022

  61. [61]

    2018, The Astrophysical Journal Supplement, 236, 26, doi: 10.3847/1538-4365/aabaf3

    Li, L., Wu, X.-F., Lei, W.-H., et al. 2018, The Astrophysical Journal Supplement, 236, 26, doi: 10.3847/1538-4365/aabaf3

    work page doi:10.3847/1538-4365/aabaf3 2018

  62. [62]

    2007, The Astrophysical Journal, 670, 565, doi: 10.1086/521870

    Liang, E.-W., Zhang, B.-B., & Zhang, B. 2007, The Astrophysical Journal, 670, 565, doi: 10.1086/521870

    work page doi:10.1086/521870 2007

  63. [63]

    T., & Wang, H

    Liu, H., Liu, C., Wang, J. T., & Wang, H. 2019, The Astrophysical Journal, 877, 121

    work page 2019

  64. [64]

    KAN: Kolmogorov-Arnold Networks

    Liu, Z., Wang, Y., Vaidya, S., et al. 2024, arXiv preprint arXiv:2404.19756

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  65. [65]

    R., Jain, L., et al

    Medsker, L. R., Jain, L., et al. 2001, Design and Applications, 5, 2

    work page 2001

  66. [66]

    Conditional Generative Adversarial Nets

    Mirza, M., & Osindero, S. 2014, Conditional Generative Adversarial Nets. https://arxiv.org/abs/1411.1784

    work page internal anchor Pith review Pith/arXiv arXiv 2014

  67. [67]

    2024, arXiv preprint arXiv:2410.13985

    Narendra, A., Dainotti, M., Sarkar, M., et al. 2024, arXiv preprint arXiv:2410.13985

    work page arXiv 2024

  68. [68]

    A., Kouveliotou, C., Grupe, D., et al

    Nousek, J. A., Kouveliotou, C., Grupe, D., et al. 2006, ApJ, 642, 389, doi: 10.1086/500724

    work page doi:10.1086/500724 2006

  69. [69]

    2007, SIAM J

    Oberst, U. 2007, SIAM J. Control and Optimization, 46, 496, doi: 10.1137/060658242 O’Brien, P. T., Willingale, R., Osborne, J., et al. 2006, The Astrophysical Journal, 647, 1213, doi: 10.1086/505457

    work page doi:10.1137/060658242 2007

  70. [70]

    Attention U-Net: Learning Where to Look for the Pancreas

    Oktay, O., Schlemper, J., Folgoc, L. L., et al. 2018, arXiv preprint arXiv:1804.03999

    work page internal anchor Pith review Pith/arXiv arXiv 2018

  71. [71]

    2000, The Astrophysical Journal, 543, 66

    Panaitescu, A., & Kumar, P. 2000, The Astrophysical Journal, 543, 66

    work page 2000

  72. [72]

    2022, IMA Journal of Mathematical Control and Information, 39, 708, doi: 10.1093/imamci/dnac005

    Pechlivanidou, G., & Karampetakis, N. 2022, IMA Journal of Mathematical Control and Information, 39, 708, doi: 10.1093/imamci/dnac005

    work page doi:10.1093/imamci/dnac005 2022

  73. [73]

    J., & Igel, C

    Perslev, M., Jensen, M., Darkner, S., Jennum, P. J., & Igel, C. 2019, Advances in neural information processing systems, 32

    work page 2019

  74. [74]

    1999, Physics Reports, 314, 575

    Piran, T. 1999, Physics Reports, 314, 575

    work page 1999

  75. [75]

    Evidence for a late-time outburst of the X-ray afterglow of GB970508 from BeppoSAX

    Piro, L., Amati, L., Antonelli, L. A., et al. 1998, Astronomy and Astrophysics, 331, L41. https://arxiv.org/abs/astro-ph/9710355

    work page internal anchor Pith review Pith/arXiv arXiv 1998

  76. [76]

    2014, The Astrophysical Journal, 783, 126

    Capozziello, S. 2014, The Astrophysical Journal, 783, 126

    work page 2014

  77. [77]

    N., et al

    Racusin, J., Liang, E., Burrows, D. N., et al. 2009, The Astrophysical Journal, 698, 43

    work page 2009

  78. [78]

    L., Liang, E

    Racusin, J. L., Liang, E. W., Burrows, D. N., et al. 2009, ApJ, 698, 43, doi: 10.1088/0004-637X/698/1/43

    work page doi:10.1088/0004-637x/698/1/43 2009

  79. [79]

    Rasmussen, C. E. 2003, in Summer school on machine learning (Springer), 63–71

    work page 2003

  80. [80]

    A., et al

    Rea, N., Gull´ on, M., Pons, J. A., et al. 2015, The Astrophysical Journal, 813, 92, doi: 10.1088/0004-637X/813/2/92

    work page doi:10.1088/0004-637x/813/2/92 2015

Showing first 80 references.