pith. sign in

arxiv: 2605.16762 · v1 · pith:SENFWQXBnew · submitted 2026-05-16 · 🌌 astro-ph.IM · astro-ph.CO

Extracting redshifts from 2D slitless spectroscopic images using deep learning for the CSST galaxy survey

Xingchen Zhou , Yan Gong , Xin Zhang , Xian-Min Meng , Haitao Miao , Run Wen , Nan Li This is my paper

Pith reviewed 2026-05-19 19:46 UTC · model grok-4.3

classification 🌌 astro-ph.IM astro-ph.CO
keywords redshift estimationslitless spectroscopydeep learningCSST surveyBayesian neural networkgalaxy redshiftscosmological surveys
0
0 comments X

The pith

A deep learning model extracts galaxy redshifts directly from 2D slitless spectroscopic images while estimating uncertainties.

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

This paper develops a method to obtain redshifts straight from 2D slitless spectral images for the CSST survey instead of first extracting 1D spectra. In slitless observations the spectrum spreads across the detector and blends with the galaxy's spatial shape, which creates calibration difficulties and reduces the quality of traditional 1D processing. The authors build a realistic mock dataset from high-resolution images and spectral templates, then train a Bayesian convolutional neural network that outputs both a redshift value and an uncertainty measure. The resulting precision reaches levels required for cosmological measurements such as baryon acoustic oscillations, and the model shows some tolerance to wavelength calibration mistakes when spatial augmentations are used during training.

Core claim

The central claim is that a Bayesian convolutional neural network implemented via Monte Carlo dropout can map CSST GV and GI band 2D slitless images to redshift estimates with σ_NMAD = 0.0104 and mean uncertainty 0.0155 for sources above SNR_GI of 1, with σ_NMAD improving to 0.0047, 0.0037 and 0.0024 at SNR_GI thresholds of 3, 5 and 10 respectively, while remaining resilient to wavelength calibration errors through spatial augmentations.

What carries the argument

Bayesian convolutional neural network with Monte Carlo dropout that directly predicts redshift and uncertainty from 2D slitless spectral images.

If this is right

  • Redshift values become available without performing the 1D spectral extraction step that is especially error-prone for slitless data.
  • Uncertainty estimates from the model can be propagated directly into cosmological analyses that rely on the CSST slitless survey.
  • The reported precision levels satisfy the accuracy targets needed for baryon acoustic oscillation measurements with CSST.
  • Spatial augmentations during training reduce sensitivity to wavelength calibration inaccuracies that commonly affect slitless spectra.

Where Pith is reading between the lines

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

  • The direct 2D-to-redshift mapping could be tested on data from other planned slitless surveys to check whether similar performance is obtained.
  • If the method works on actual observations it might shorten the time required to produce large redshift catalogs from wide-field imaging spectroscopy.
  • Further experiments that vary the noise properties in the mock data could identify the conditions under which the precision degrades.

Load-bearing premise

The mock images and spectra created from HSC-SSP and DESI data correctly reproduce the noise, point-spread function and wavelength calibration behavior of actual CSST GV and GI observations.

What would settle it

Running the trained network on real CSST slitless data and comparing the predicted redshifts and uncertainties against independent spectroscopic measurements or traditional 1D extractions on the same objects would test whether the reported precision is achieved in practice.

Figures

Figures reproduced from arXiv: 2605.16762 by Haitao Miao, Nan Li, Run Wen, Xian-Min Meng, Xingchen Zhou, Xin Zhang, Yan Gong.

Figure 2
Figure 2. Figure 2: The density plot of SNRGI with respect to ztrue. The color bar indicates the number of sources within each pixel. any negative flux values are artificially set to 0. The distribution of SNRGI for the samples is presented in [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 1
Figure 1. Figure 1: The distribution of SNR at GI band, SNRGI . The red, green, blue and yellow lines indicate the SNR thresholds of 1, 3, 5 and 10, respectively. The number of sources exceeding each threshold is provided in the legend. In this work, we restrict our analysis to sources with a mean SNR calculated by 1D spectra at the GI band with SNRGI ≥ 1. The mean SNR is defined as SNR = 1 n X i fi ei (3) where fi and ei den… view at source ↗
Figure 3
Figure 3. Figure 3: Four representative examples of mock slitless spectra at various redshifts and SNRGI . For each example, the leftmost panel displays the ideal spatial image, while the two upper right panels illustrate the corresponding 2D spectral images in the GV and GI bands. The lower panel compares the intrinsic SED with the extracted 1D spectrum [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Redshift distributions of the training (blue) and testing (orange) datasets. The testing set is explicitly sam￾pled to follow the anticipated distribution of CSST slitless spectroscopic survey. passes through the network. Due to the active dropout layers in testing stage, each individual forward pass t yields a predicted mean redshift, µt, and a predicted aleatoric variance, σ 2 t . The final, combined red… view at source ↗
Figure 5
Figure 5. Figure 5: The density plot of the spec-z predictions from the deterministic pre-training network. The color bar indi￾cates the number of sources within each pixel. Overall, this base network achieves a precision σNMAD = 0.0097 for all sources with SNRGI ≥ 1.0 [PITH_FULL_IMAGE:figures/full_fig_p007_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Confidence versus coverage plot used to evaluate the statistical reliability of the predicted uncertainties. The gray dashed, blue, and red curves illustrate the ideal, pre– calibration, and post-calibration scenarios, respectively. The derived temperature scaling factor of T = 1.400 indicates that the raw uncertainties were initially under-estimated. uate the deterministic pre-trained network [PITH_FULL_… view at source ↗
Figure 7
Figure 7. Figure 7: Predicted redshift zpred versus true redshift ztrue evaluated across four GI-band SNR thresholds SNRGI ≥ 1.0, 3.0, 5.0, and 10.0. The error-bars indicate the calibrated total uncertainties. Each panel reports the corresponding predictive precision σNMAD and the mean normalized uncertainty ⟨E/(1 + ztrue)⟩. To maintain visual clarity, a random subsample of 5,000 sources is shown in the first three panels, wh… view at source ↗
Figure 9
Figure 9. Figure 9 [PITH_FULL_IMAGE:figures/full_fig_p010_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: Redshift distributions for the BGS (orange), LRG (green), and ELG (blue) source populations, classified according to the main DESI target selection criteria. The total number of sources for each target class is provided in the legend. lustrated in [PITH_FULL_IMAGE:figures/full_fig_p011_10.png] view at source ↗
read the original abstract

Wide-field slitless spectroscopic galaxy surveys, such as the one performed by the upcoming Chinese Space Station Survey Telescope (CSST), are crucial for precision cosmology but present formidable data analysis challenges. Because spectra are dispersed directly onto the detector, they are convolved with the 2-dimensional (2D) spatial morphology, which complicates wavelength calibration and consequently degrades the fidelity of subsequent 1-dimensional (1D) spectral extraction. To overcome these limitations, we present a deep learning framework that extracts redshifts directly from 2D slitless spectral images, bypassing 1D extraction entirely. We construct a realistic mock dataset for the CSST $GV$ and $GI$ band using high-resolution images from HSC-SSP PDR3 and spectral energy distributions (SEDs) from DESI DR1. A Bayesian convolutional neural network implemented by Monte Carlo dropout is employed to map the 2D spectral images to redshift estimations while simultaneously quantifying uncertainties. We find that our model can achieve a precision $\sigma_{\rm NMAD}=0.0104$ and mean uncertainty $\langle E / (1 + z_{{\rm true}}) \rangle=0.0155$ for sources with ${\rm SNR}_{GI}\geq1$. For sources with ${\rm SNR}_{GI}$ higher than 3.0, 5.0 and 10.0, $\sigma_{\rm NMAD}$ can achieve 0.0047, 0.0037 and 0.0024 respectively, matching the redshift precision requirements for studies such as BAO using the CSST slitless spectroscopic surveys. Furthermore, by utilizing spatial augmentations, the network demonstrates resilience to wavelength calibration errors. This work provides a novel and robust pathway for data analysis of next-generation slitless spectroscopic galaxy surveys.

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

1 major / 2 minor

Summary. The paper introduces a Bayesian convolutional neural network with Monte Carlo dropout to extract galaxy redshifts directly from 2D slitless spectroscopic images for the upcoming CSST survey, avoiding traditional 1D extraction. A mock dataset is constructed by injecting DESI DR1 SEDs into HSC-SSP PDR3 images and applying a CSST forward model for the GV and GI bands. On this mock, the model achieves σ_NMAD = 0.0104 (with mean uncertainty 0.0155) for SNR_GI ≥ 1, tightening to 0.0024 at SNR_GI ≥ 10, and the authors state that these values meet BAO precision requirements; spatial augmentations are used to demonstrate robustness to wavelength calibration errors.

Significance. If the mock faithfully captures CSST instrument properties, the direct 2D approach could offer a practical solution to morphology-induced wavelength calibration issues in slitless spectroscopy, with built-in uncertainty estimates that are valuable for cosmological analyses. The reported tightening of precision with SNR cuts and the augmentation-based robustness test are concrete strengths that would support adoption in survey pipelines if validated.

major comments (1)
  1. [Dataset construction] Dataset construction (abstract and methods paragraph): The headline claim that σ_NMAD values 'match the redshift precision requirements for studies such as BAO using the CSST slitless spectroscopic surveys' is load-bearing on the mock dataset reproducing the joint statistics of 2D morphology convolution, instrument PSF, read noise, and wavelength solution at the level relevant to redshift extraction. No quantitative validation (e.g., line-spread-function width histograms, power-spectrum residuals, or end-to-end comparison to an independent CSST simulator) is supplied; only an assertion of realism is given. This directly affects whether the quoted numbers can be taken as indicative of real-survey performance.
minor comments (2)
  1. [Results] Results: The reported NMAD and uncertainty metrics lack error bars or bootstrap uncertainties, which would help assess the statistical significance of the quoted improvements with SNR cuts.
  2. [Methods] Methods: The precise network architecture (number of layers, filter sizes, dropout rate schedule) and training details (loss function, optimizer, data split) should be expanded for reproducibility.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for their careful reading and constructive feedback on our manuscript. We address the single major comment below and have revised the manuscript to incorporate additional quantitative validation of the mock dataset.

read point-by-point responses

  1. Referee: Dataset construction (abstract and methods paragraph): The headline claim that σ_NMAD values 'match the redshift precision requirements for studies such as BAO using the CSST slitless spectroscopic surveys' is load-bearing on the mock dataset reproducing the joint statistics of 2D morphology convolution, instrument PSF, read noise, and wavelength solution at the level relevant to redshift extraction. No quantitative validation (e.g., line-spread-function width histograms, power-spectrum residuals, or end-to-end comparison to an independent CSST simulator) is supplied; only an assertion of realism is given. This directly affects whether the quoted numbers can be taken as indicative of real-survey performance.

    Authors: We appreciate the referee highlighting the need for explicit validation of the mock. The dataset is built from real HSC-SSP PDR3 images (providing observed 2D morphologies and noise properties) and DESI DR1 SEDs, passed through a CSST-specific forward model that includes the instrument PSF, read noise, and wavelength dispersion for the GV and GI bands. We agree that quantitative checks strengthen the claim that the reported σ_NMAD values are indicative of survey performance. In the revised manuscript we have added a new subsection to the Methods section containing (i) histograms of simulated line-spread-function widths compared against the expected CSST values, (ii) power-spectrum residuals between the mock images and the input HSC data after convolution, and (iii) a brief end-to-end comparison against an independent, simplified CSST simulator. These additions demonstrate that the joint statistics relevant to redshift extraction are reproduced at the level needed to support the quoted precision and the BAO-requirement statement. revision: yes

Circularity Check

0 steps flagged

No significant circularity in derivation chain

full rationale

The paper reports empirical performance metrics from training and testing a Bayesian CNN on a mock dataset built from external HSC-SSP and DESI sources. These σ_NMAD and uncertainty values are direct evaluation outputs on held-out mock images and do not reduce by the paper's own equations or definitions to fitted inputs or self-citation loops. No load-bearing self-citations, uniqueness theorems, or ansatzes from prior author work are invoked to justify the central claims. The methodology is a standard data-driven pipeline whose results remain independent of the reported numbers themselves.

Axiom & Free-Parameter Ledger

1 free parameters · 1 axioms · 0 invented entities

The central claim rests on the assumption that the HSC+DESI mock images are statistically representative of CSST observations and that the network generalizes beyond the training distribution. No new physical entities are postulated. The only free parameters are the internal weights of the neural network, which are fitted during training rather than chosen by hand for the final result.

free parameters (1)
  • neural network weights
    Learned during supervised training on the mock dataset; not a scientific free parameter but the source of the model's mapping.
axioms (1)
  • domain assumption The noise and PSF properties in the HSC-SSP PDR3 images combined with DESI DR1 SEDs accurately simulate CSST GV/GI observations.
    Invoked when constructing the training and test sets described in the abstract.

pith-pipeline@v0.9.0 · 5877 in / 1607 out tokens · 48713 ms · 2026-05-19T19:46:06.401661+00:00 · methodology

discussion (0)

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

Reference graph

Works this paper leans on

42 extracted references · 42 canonical work pages · 12 internal anchors

  1. [1]

    The Hyper Suprime-Cam SSP Survey: Overview and Survey Design

    Aihara, H., Arimoto, N., Armstrong, R., et al. 2018, PASJ, 70, S4, doi: 10.1093/pasj/psx066

    work page internal anchor Pith review doi:10.1093/pasj/psx066 2018

  2. [2]

    2022, PASJ, 74, 247, doi: 10.1093/pasj/psab122

    Aihara, H., AlSayyad, Y., Ando, M., et al. 2022, PASJ, 74, 247, doi: 10.1093/pasj/psab122

    work page doi:10.1093/pasj/psab122 2022

  3. [3]

    2013, Living Reviews in Relativity, 16, 6, doi: 10.12942/lrr-2013-6

    Amendola, L., Appleby, S., Bacon, D., et al. 2013, Living Reviews in Relativity, 16, 6, doi: 10.12942/lrr-2013-6

    work page doi:10.12942/lrr-2013-6 2013

  4. [4]

    Baryon Acoustic Oscillations

    Bassett, B., & Hlozek, R. 2010, in Dark Energy, ed. P. Ruiz-Lapuente, 246, doi: 10.48550/arXiv.0910.5224

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.0910.5224 2010

  5. [5]

    R., Lin, H., Lupton, R

    Blanton, M. R., Lin, H., Lupton, R. H., et al. 2003, AJ, 125, 2276, doi: 10.1086/344761

    work page doi:10.1086/344761 2003

  6. [6]

    B., van Dokkum, P

    Brammer, G. B., van Dokkum, P. G., Franx, M., et al. 2012, ApJS, 200, 13, doi: 10.1088/0067-0049/200/2/13

    work page doi:10.1088/0067-0049/200/2/13 2012

  7. [7]

    2016, ARA&A, 54, 597, doi: 10.1146/annurev-astro-082214-122432

    Cappellari, M. 2016, ARA&A, 54, 597, doi: 10.1146/annurev-astro-082214-122432

    work page doi:10.1146/annurev-astro-082214-122432 2016

  8. [8]

    The DESI Experiment Part I: Science,Targeting, and Survey Design

    Conroy, C. 2013, ARA&A, 51, 393, doi: 10.1146/annurev-astro-082812-141017 13 CSST Collaboration, Gong, Y., Miao, H., et al. 2026, Science China Physics, Mechanics, and Astronomy, 69, 239501, doi: 10.1007/s11433-025-2809-0 DESI Collaboration, Aghamousa, A., Aguilar, J., et al. 2016, arXiv e-prints, arXiv:1611.00036, doi: 10.48550/arXiv.1611.00036 DESI Coll...

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.1146/annurev-astro-082812-141017 2013

  9. [9]

    and Dambre, Joni , year=

    Dieleman, S., Willett, K. W., & Dambre, J. 2015, MNRAS, 450, 1441, doi: 10.1093/mnras/stv632

    work page doi:10.1093/mnras/stv632 2015

  10. [10]

    Farahani, A., Pourshojae, B., Rasheed, K., & Arabnia, H. R. 2021, arXiv e-prints, arXiv:2104.02144, doi: 10.48550/arXiv.2104.02144

    work page doi:10.48550/arxiv.2104.02144 2021

  11. [11]

    Ferreira, P., & Reis, R. R. R. 2025, Brazilian Journal of Physics, 55, 33, doi: 10.1007/s13538-024-01669-7

    work page doi:10.1007/s13538-024-01669-7 2025

  12. [12]

    Dropout as a Bayesian Approximation: Representing Model Uncertainty in Deep Learning

    Gal, Y., & Ghahramani, Z. 2015, arXiv e-prints, arXiv:1506.02142, doi: 10.48550/arXiv.1506.02142

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.1506.02142 2015

  13. [13]

    Gallazzi, A., Charlot, S., Brinchmann, J., White, S. D. M., & Tremonti, C. A. 2005, MNRAS, 362, 41, doi: 10.1111/j.1365-2966.2005.09321.x

    work page doi:10.1111/j.1365-2966.2005.09321.x 2005

  14. [14]

    Bayesian Neural Networks: An Introduction and Survey

    Goan, E., & Fookes, C. 2020, arXiv e-prints, arXiv:2006.12024, doi: 10.48550/arXiv.2006.12024

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2006.12024 2020

  15. [15]

    2019, ApJ, 883, 203, doi: 10.3847/1538-4357/ab391e

    Gong, Y., Liu, X., Cao, Y., et al. 2019, ApJ, 883, 203, doi: 10.3847/1538-4357/ab391e

    work page doi:10.3847/1538-4357/ab391e 2019

  16. [16]

    2025, Science China

    Gong, Y., Miao, H., Zhou, X., et al. 2025, Science China

    work page 2025

  17. [17]

    Physics, Mechanics, and Astronomy, 68, 280402, doi: 10.1007/s11433-025-2646-2

    work page doi:10.1007/s11433-025-2646-2

  18. [18]

    J., Ruiz-Macias, O., et al

    Hahn, C., Wilson, M. J., Ruiz-Macias, O., et al. 2023, AJ, 165, 253, doi: 10.3847/1538-3881/accff8

    work page doi:10.3847/1538-3881/accff8 2023

  19. [19]

    Hamilton, A. J. S. 1998, in Astrophysics and Space Science

    work page 1998

  20. [20]

    231, The Evolving Universe, ed

    Library, Vol. 231, The Evolving Universe, ed. D. Hamilton, 185, doi: 10.1007/978-94-011-4960-0 17

    work page doi:10.1007/978-94-011-4960-0

  21. [21]

    Deep Residual Learning for Image Recognition

    He, K., Zhang, X., Ren, S., & Sun, J. 2015, Deep Residual Learning for Image Recognition, https://arxiv.org/abs/1512.03385

    work page internal anchor Pith review Pith/arXiv arXiv 2015

  22. [22]

    Gaussian Error Linear Units (GELUs)

    Hendrycks, D., & Gimpel, K. 2016, arXiv e-prints, arXiv:1606.08415, doi: 10.48550/arXiv.1606.08415

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.1606.08415 2016

  23. [23]

    D., Perreault Levasseur, L., & Marshall, P

    Hezaveh, Y. D., Perreault Levasseur, L., & Marshall, P. J. 2017, Nature, 548, 555, doi: 10.1038/nature23463 K¨ ummel, M., Walsh, J. R., Pirzkal, N., Kuntschner, H., &

    work page doi:10.1038/nature23463 2017

  24. [24]

    2009, PASP, 121, 59, doi: 10.1086/596715

    Pasquali, A. 2009, PASP, 121, 59, doi: 10.1086/596715

    work page doi:10.1086/596715 2009

  25. [25]

    Euclid Definition Study Report

    Laureijs, R., Amiaux, J., Arduini, S., et al. 2011, arXiv e-prints, arXiv:1110.3193, doi: 10.48550/arXiv.1110.3193

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.1110.3193 2011

  26. [26]

    Nature, 521, 436 –444, https://doi.org/10.1038/nature14539

    Lecun, Y., Bengio, Y., & Hinton, G. 2015, Nature, 521, 436, doi: 10.1038/nature14539

    work page doi:10.1038/nature14539 2015

  27. [27]

    Decoupled Weight Decay Regularization

    Loshchilov, I., & Hutter, F. 2017, arXiv e-prints, arXiv:1711.05101, doi: 10.48550/arXiv.1711.05101

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.1711.05101 2017

  28. [28]

    2014, ARA&A, 52, 415, doi: 10.1146/annurev-astro-081811-125615

    Madau, P., & Dickinson, M. 2014, ARA&A, 52, 415, doi: 10.1146/annurev-astro-081811-125615

    work page internal anchor Pith review doi:10.1146/annurev-astro-081811-125615 2014

  29. [29]

    2024, MNRAS, 531, 3991, doi: 10.1093/mnras/stae1370

    Miao, H., Gong, Y., Chen, X., et al. 2024, MNRAS, 531, 3991, doi: 10.1093/mnras/stae1370

    work page doi:10.1093/mnras/stae1370 2024

  30. [30]

    G., Brammer, G

    Momcheva, I. G., Brammer, G. B., van Dokkum, P. G., et al. 2016, ApJS, 225, 27, doi: 10.3847/0067-0049/225/2/27

    work page doi:10.3847/0067-0049/225/2/27 2016

  31. [31]

    J., Bennett , C., et al

    Mosby, G., Rauscher, B. J., Bennett, C., et al. 2020, Journal of Astronomical Telescopes, Instruments, and Systems, 6, 046001, doi: 10.1117/1.JATIS.6.4.046001

    work page doi:10.1117/1.jatis.6.4.046001 2020

  32. [32]

    Bayesian Neural Networks

    Mullachery, V., Khera, A., & Husain, A. 2018, arXiv e-prints, arXiv:1801.07710, doi: 10.48550/arXiv.1801.07710

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.1801.07710 2018

  33. [33]

    A., & Gruen, D

    Newman, J. A., & Gruen, D. 2022, ARA&A, 60, 363, doi: 10.1146/annurev-astro-032122-014611

    work page doi:10.1146/annurev-astro-032122-014611 2022

  34. [34]

    Papadakis, C., Filandrianos, G., Dimitriou, A., Lymperaiou, M., Thomas, K., Stamou, G., 2025

    Pan, S. J., & Yang, Q. 2010, IEEE Transactions on Knowledge and Data Engineering, 22, 1345, doi: 10.1109/TKDE.2009.191

    work page doi:10.1109/tkde.2009.191 2010

  35. [35]

    A., et al

    Raichoor, A., Moustakas, J., Newman, J. A., et al. 2023, AJ, 165, 126, doi: 10.3847/1538-3881/acb213

    work page doi:10.3847/1538-3881/acb213 2023

  36. [36]

    Richardson, W. H. 1972, J. Opt. Soc. Am., 62, 55, doi: 10.1364/JOSA.62.000055

    work page doi:10.1364/josa.62.000055 1972

  37. [37]

    2019, Nature Astronomy, 3, 212, doi: 10.1038/s41550-018-0478-0

    Salvato, M., Ilbert, O., & Hoyle, B. 2019, Nature Astronomy, 3, 212, doi: 10.1038/s41550-018-0478-0

    work page doi:10.1038/s41550-018-0478-0 2019

  38. [38]

    H., Mortonson, M

    Weinberg, D. H., Mortonson, M. J., Eisenstein, D. J., et al. 2013, PhR, 530, 87, doi: 10.1016/j.physrep.2013.05.001

    work page doi:10.1016/j.physrep.2013.05.001 2013

  39. [39]

    Weiner, B. J. 2012, arXiv e-prints, arXiv:1209.1405, doi: 10.48550/arXiv.1209.1405

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.1209.1405 2012

  40. [40]

    A., et al

    Zhou, R., Dey, B., Newman, J. A., et al. 2023, AJ, 165, 58, doi: 10.3847/1538-3881/aca5fb

    work page doi:10.3847/1538-3881/aca5fb 2023

  41. [41]

    2024, ApJ, 977, 69, doi: 10.3847/1538-4357/ad8bbf

    Zhou, X., Gong, Y., Zhang, X., et al. 2024, ApJ, 977, 69, doi: 10.3847/1538-4357/ad8bbf

    work page doi:10.3847/1538-4357/ad8bbf 2024

  42. [42]

    2021, IEEE Proceedings, 109, 43, doi: 10.1109/JPROC.2020.3004555

    Zhuang, F., Qi, Z., Duan, K., et al. 2021, IEEE Proceedings, 109, 43, doi: 10.1109/JPROC.2020.3004555

    work page doi:10.1109/jproc.2020.3004555 2021