Pith. sign in

REVIEW 4 major objections 6 minor 19 references

Hybrid-z: Enhancing Kilo-Degree Survey bright galaxy sample photometric redshifts with deep learning

T0 review · 4 major / 6 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read A deep-learning model that combines four-band galaxy images with nine-band magnitudes cuts photometric-redshift scatter for the KiDS-Bright sample by about 20% over the previous ANNz2 results, from SMAD($\Delta z$) $\approx 0.018$ to…

desk verdict Genuine 20% scatter improvement over ANNz2 for KiDS-Bright, honestly documented, but the public catalog comes from a retrained model that never got a spectroscopic validation. read the letter →

arxiv 2501.01942 v2 pith:3GKLUO42 submitted 2025-01-03 astro-ph.CO

Anjitha John William , Priyanka Jalan , Maciej Bilicki , Wojciech A. Hellwing , Hareesh Thuruthipilly , Szymon J. Nakoneczny This is my paper · ORCID
classification astro-ph.CO
keywords photometricredshiftsdeeplearningconvolutionalneuralnetworksKiDS-BrightGAMAspectroscopygalaxysurveyscosmicvariance
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 builds a hybrid deep-learning model, Hybrid-z, that estimates photometric redshifts for the bright, flux-limited KiDS-Bright galaxy sample by feeding four-band KiDS images through a convolutional network and nine-band KiDS+VIKING magnitudes through an ordinary neural network, then concatenating the two. Tested on GAMA spectroscopy, the model reaches a scatter of SMAD($\Delta z$) $\approx 0.014(1+z)$ on the clean sample, about 20% smaller than the earlier ANNz2 value of $\approx 0.018(1+z)$, while keeping the mean bias at a few times $10^{-4}$. The improvement is largest for blue galaxies, whose scatter drops by about 22%, consistent with image morphology carrying information that magnitudes alone do not. For the full 1.2-million-galaxy catalog the authors retrain on a smoothed subsample of GAMA redshifts to prevent cosmic-variance features in the training sky from being imprinted on the predicted redshift distribution, releasing this final catalog alongside the paper. If correct, the work shows that adding imaging to magnitude-based photo-$z$ networks gives a material precision gain for the bright foreground galaxies used in lensing and clustering analyses.

What carries the argument

The load-bearing mechanism is the Hybrid-z architecture and its training procedure. A convolutional network built from inception modules, using parallel $1\times1$, $3\times3$, and $5\times5$ convolutions with average pooling, extracts multi-scale features from $36\times36$-pixel, four-band ($ugri$) KiDS cutouts; an ordinary fully-connected network processes nine standardized magnitudes; and the two outputs are depth-concatenated before the final dense layers, an architectural choice that lets image morphology directly influence the redshift prediction. The second mechanism is the smoothing algorithm applied for the final catalog: iteratively trimming spikes from the GAMA training redshift histogram to produce a more uniform $dz$ distribution, which the authors use to retrain the model and thereby suppress redshift-focusing artifacts in the full-sample prediction.

What would settle it

Take the released Hybrid-z catalog and compare its photo-$z$s with an independent spectroscopic sample covering the faint part of KiDS-Bright ($r$ between 19.5 and 20 mag) that was not used in training, for instance DESI DR1 or a dedicated follow-up; if the SMAD($\Delta z$) there exceeds about $0.018(1+z)$ or the mean bias grows beyond a few times $10^{-3}$, the claim that the full 1.2-million-galaxy catalog achieves the reported precision would be refuted.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central discovery is that a deep network processing galaxy images together with nine-band magnitudes yields photometric redshifts for KiDS-Bright that are substantially more precise than the previous feature-only ANNz2 estimates, while remaining equally unbiased. Specifically, on the fiducial clean GAMA test set, Hybrid-z achieves SMAD($\Delta z$) $\approx 0.014$ compared with $\approx 0.018$ for ANNz2, a 20% reduction, with the mean of $\delta z$ at most a few times $10^{-4}$; the same pattern holds across all external spectroscopic cross-checks, with scatter reductions of 17%–24% and, in the blind samples (GAMA G23 and 2dFLenS) that are disjoint from the training fields, reductions of 17% and 20%. The paper additionally reports that applying the model trained on the full GAMA data to the entire KiDS-Bright sample imprints GAMA's large-scale-structure features, such as a dip near $z \sim 0.25$ and a peak near $z \sim 0.38$, into the predicted redshift distribution, and that training on a smoothed version of the GAMA redshift histogram removes these artifacts while, per the authors, preserving the color–redshift relation.

Load-bearing premise

The load-bearing premise is that smoothing the GAMA training redshifts removes cosmic-variance artifacts without changing the mapping from photometry to redshift, and it also assumes the model generalizes to the faintest KiDS-Bright galaxies where GAMA is incomplete, while the final model is never validated against spectroscopy.

Editorial extensions

If this is right

  • The KiDS-Bright photo-$z$ catalog recommended for science gains about 20% in scatter compared with the ANNz2 version, reducing redshift-error systematics in galaxy-galaxy lensing and clustering analyses that use these galaxies as a foreground.
  • Blue, typically spiral galaxies receive the largest improvement (about 22% scatter reduction), so samples with a high blue fraction will see the biggest gains in photo-$z$ precision.
  • Because the model performs consistently better across several independent spectroscopic samples, including two fully blind southern fields, the improvement is not an artifact of the particular GAMA equatorial test split.
  • The released full-sample catalog is trained on a smoothed redshift distribution, so its $dN/dz$ no longer mimics GAMA's cosmic-variance features, which is the relevant quantity for tomographic and clustering applications.

Reading between the lines

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

  • The headline scatter of $0.014(1+z)$ is measured only for the unsmoothed GAMA-trained model; the released catalog's retrained model is never validated against spectroscopy, so a direct external check of the released redshifts could still reveal that the smoothing changes the photometry-to-redshift mapping in ways the reported statistics do not cover.
  • The larger gains for blue galaxies suggest that morphological information from the images is a major driver of the improvement; an ablation study comparing image-only, magnitude-only, and hybrid models on matched training would quantify this contribution and is an obvious next step.
  • The same hybrid architecture could plausibly transfer to other bright, low-redshift surveys such as SDSS or future LSST data, with the main risk being covariate shift at the faint end where the training sample is incomplete (roughly $r > 19.5$ for GAMA).
  • The redshift-focusing artifacts reveal that the network is sensitive to the training $dN/dz$; rather than only smoothing GAMA, training on a wider-area spectroscopic sample such as DESI DR1 would remove the cosmic-variance imprint at the source.
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

4 major / 6 minor

Summary. The paper presents Hybrid-z, a deep-learning photometric redshift model for the KiDS-Bright DR4 sample (r < 20 mag). The model combines a convolutional network operating on four-band KiDS image cutouts with a fully connected network operating on nine-band KiDS+VIKING magnitudes, and is trained on spectroscopic redshifts from the GAMA equatorial fields. Against a held-out GAMA test sample, the authors report SMAD(Δz) ≈ 0.014(1+z) for the clean sample, a roughly 20% improvement over the previous ANNz2 results (≈ 0.018), with comparable small mean residuals; they also validate the model on the external samples SDSS, 2dFGRS, GAMA G23, and 2dFLenS, finding consistently lower scatter than ANNz2. For the final catalog of ~1.2 million galaxies, the network is retrained on a GAMA subsample whose redshift distribution has been smoothed to suppress cosmic-variance features, and the resulting photo-z catalog is publicly released.

Significance. If the headline result holds, this is a useful and practical improvement for a widely used low-redshift foreground sample: it demonstrates that image-plus-magnitude deep learning beats feature-only ANNz2 in the bright, well-resolved regime, and it provides a public catalog. The paper's main methodological strengths are genuine: the internal comparison uses a held-out split, the external validation includes two spectroscopic samples (GAMA G23 and 2dFLenS) that are spatially disjoint from the training data, and the improvement over ANNz2 is reported across multiple independent samples with consistent statistics. The central weakness is that the released catalog is produced by a retrained, smoothed model whose accuracy is never spectroscopically validated, so the paper does not currently establish that the public catalog has the measured 0.014(1+z) scatter.

major comments (4)
  1. [Section 5.3; Tables 1 and 2] The headline precision, SMAD(Δz) ≈ 0.014 for the clean sample, and all external-sample statistics in Tables 1 and 2 are obtained with the model trained on the full GAMA-equatorial sample. The released 1.2M-galaxy catalog is instead generated by a model retrained on the smoothed 118k-galaxy subsample described in Section 5.3. No spectroscopic validation of this retrained model is presented; the only displayed output of the smoothed model is the dN/dzphot distribution in Fig. 7b. The paper therefore does not establish that the released catalog has the claimed scatter or the claimed near-zero mean residuals. Please validate the smoothed model on a held-out GAMA test set and, ideally, on the blind external samples G23 and 2dFLenS, reporting the same statistics as in Tables 1 and 2. If that is not possible, the claims and the catalog description should be restricted to the unsmoothed model.
  2. [Abstract; Table 2] The abstract's statement of 'negligible mean residuals of O(10^-4)' is not supported by the external validation. Table 2 lists mean rescaled biases of -0.0031 for GAMA-Equatorial, -0.0023 for SDSS DR16, -0.0025 for 2dFLenS, and -0.0017 for GAMA G23. These offsets are small relative to the scatter, but they are of order 10^-3, not 10^-4. The O(10^-4) level applies to the internal GAMA test sample in Table 1. Please rephrase the abstract and Section 6 to attribute the O(10^-4) claim only to the internal test set and to state explicitly that external samples show systematic offsets of a few times 10^-3.
  3. [Section 5.3, smoothing procedure] The smoothing procedure subsamples the training set in redshift to flatten dN/dzspec, changing the training prior p(z). For a regression model with intrinsic scatter, the optimal conditional mean prediction E[z|x] depends on that prior, so preserving the color-redshift relation p(color|z) does not guarantee that p(z|color) or the conditional photo-z mapping is preserved. The paper's assertion that smoothing does not affect the color-redshift relation is therefore not a substitute for end-to-end spectroscopic validation of the retrained model. This concern is reinforced by the paper's own observation that a peak at z_phot ≈ 0.38 persists in the smoothed model (Section 5.3, Fig. 7b), indicating that the mitigation is incomplete. Please provide bias and scatter as functions of true redshift, magnitude, and color for the smoothed model on spectroscopic data, and discuss how the smoothing shifts the conditional predictions relative to the validated unsmoothed model.
  4. [Section 5.3, faint end] The paper acknowledges that GAMA becomes incomplete at r ≳ 19.5 mag and that this causes covariate shift for the faintest KiDS-Bright galaxies, yet the magnitude dependence of the released (smoothed) model is never quantified. Fig. 5c shows growing scatter at r near 20 for the unsmoothed model, but the released catalog uses the smoothed model, whose faint-end behavior may differ. Given that the catalog is flux-limited to r < 20 and contains over a million galaxies, the faint-end extrapolation is a load-bearing uncertainty for the public product. Please report magnitude-binned statistics for the smoothed model on spectroscopic data, or explicitly quantify the resulting systematic uncertainty in the released catalog.
minor comments (6)
  1. [Equations (3)-(4)] The text states that z_i is the predicted value and \hat{z}_i is the true value, which is the opposite of the standard convention used elsewhere in the paper; please swap the notation or the definitions for consistency.
  2. [Equation (2)] The symbol m is used both for the magnitude and for the mean in the standardization formula; please use a different symbol for the mean, such as \mu_m, to avoid ambiguity.
  3. [Section 3.2] The paper says about 125k galaxies are used for training, but the stated 70:15:15 split of the ~173k equatorial galaxies gives roughly 121k; please reconcile these numbers.
  4. [Section 5.1] The phrase 'we would such as to emphasize' should read 'we would like to emphasize'.
  5. [Fig. 2 caption] The caption contains a typo: 'Input (36,36,,4)' should be 'Input (36,36,4)'.
  6. [Fig. 5 caption] The caption calls the GAMA test sample 'blind', but it is a random split from the same GAMA-equatorial distribution used for training; the genuinely blind samples are GAMA G23 and 2dFLenS. Please adjust the wording to avoid overstating independence.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: Hybrid-z's headline improvement over ANNz2 is measured on held-out GAMA and disjoint external spectroscopic samples; the Sect. 5.3 smoothing caveat is a validation gap, not an identity.

full rationale

The derivation chain for the main claim is an out-of-sample supervised regression. Hybrid-z is trained on GAMA equatorial labels (images plus nine-band magnitudes), and the reported SMAD(Δz) ≈ 0.014(1+z) versus ANNz2's ≈ 0.018 is computed on a held-out GAMA test set (Table 1) and, for the blind cases, on GAMA G23 and 2dFLenS, which are disjoint from the training fields (Table 2). These statistics are not part of the loss function or of any fitted parameter, so the 20% improvement does not reduce to the training input by construction. The ANNz2 baseline is a previously published catalog (B21); citing it for comparison is normal scientific practice, not load-bearing self-citation. The only in-scope weakness is in Sect. 5.3: the released 1.2M-galaxy catalog is produced by a model retrained on a z-smoothed GAMA subsample, and that retrained model is never validated against spectroscopy; the headline scatter is quoted for the unsmoothed model. This is a genuine validity risk (the smooth dN/dz is in part a designed consequence of the label reweighting, and covariate shift at r > 19.5 mag remains unquantified), but it does not make the main improvement claim equivalent to its inputs. No step in the paper fits the enumerated circularity patterns, so the score is 0.

Assumptions & free parameters 7 free parameters · 6 assumptions · 0 invented entities

The central claims rest on the supervised training setup: GAMA labels, preprocessing choices, and a set of hand-tuned hyperparameters. No new physical entities are introduced. The smoothing procedure in Sect. 5.3 is the least controlled element: it is a post hoc reweighting with unspecified threshold parameters, and its effect on the color-redshift relation is asserted, not measured.

free parameters (7)
  • learning_rate = 1e-4
    Selected by iterative tests over [1e-5, 1e-2]; affects optimization and final model (Sect. 4).
  • huber_loss_delta = 0.001
    Chosen from tests over [1e-5, 0.01] as giving lowest Huber loss, MSE, and MAE; controls outlier sensitivity (Sect. 3.5).
  • cutout_size = 7.2 arcsec (36x36 px)
    Adopted after trying 8 and 20.2 arcsec; affects image content and model input (Sect. 2.1).
  • num_inception_modules = 4
    Tested 3 to 5 and chose 4; architecture hyperparameter (Sect. 4).
  • augmentation_shift = 5% width/height shifts and flips
    Selected by experimentation; increases training set to about 484k objects (Sect. 3.3).
  • smoothing_threshold_and_binning = not specified
    Iterative bin-count adjustment to flatten GAMA dN/dz; threshold and binning are unstated, and the resulting 118k (66%) subsample is a fitted construction (Sect. 5.3).
  • sigmoid_output_cap = z in (0,1)
    Output activation restricts predictions below unity, sacrificing the rare z>1 objects; architectural choice (Sect. 4).
assumptions (6)
  • domain assumption GAMA spectroscopic redshifts with NQ>=3 are accurate labels for training and testing.
    Sect. 2.3; the whole supervised pipeline assumes these labels are correct and unbiased.
  • domain assumption The KiDS-GAMA crossmatch with 1 arcsec radius correctly identifies galaxy counterparts.
    Sect. 2.3; wrong matches would degrade the training labels and test statistics.
  • domain assumption GAMA equatorial is representative of the KiDS-Bright sample across the full magnitude range.
    Sect. 5.3 acknowledges incompleteness at r>19.5 mag, creating a covariate shift at the faint end; the model extrapolates there.
  • ad hoc to paper Smoothing the training redshift distribution does not alter the color-redshift relation.
    Sect. 5.3 claims this but does not demonstrate it; if false, the final catalog redshifts are biased in an untested way.
  • standard math Pixel normalization (Eq. 1) and magnitude standardization (Eq. 2) preserve information needed for redshift estimation.
    Preprocessing choices; per-galaxy scaling could in principle discard overall flux information, though magnitudes provide that separately.
  • domain assumption The redshift range is effectively limited to z<1 because of the r<20 mag flux limit.
    Sect. 4; objects beyond z=1 are rare, but the sigmoid cap precludes predicting them, which could bias the very high-redshift tail.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Hybrid-z: Enhancing Kilo-Degree Survey bright galaxy sample photometric redshifts with deep learning." pith.science (2026). https://pith.science/paper/3GKLUO42

@misc{pith2026250101942,
  author       = {Pith},
  title        = {Pith review of: Hybrid-z: Enhancing Kilo-Degree Survey bright galaxy sample photometric redshifts with deep learning},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/3GKLUO42}},
  note         = {Machine review of arXiv:2501.01942}
}
abstract

We employ deep learning (DL) to improve photometric redshifts (photo-$z$s) in the Kilo-Degree Survey Data Release 4 Bright galaxy sample (KiDS-Bright DR4). This dataset, used as a foreground for KiDS lensing and clustering studies, is flux-limited to $r<20$ mag with mean $z=0.23$ and covers 1000 deg$^2$. Its photo-$z$s were previously derived with artificial neural networks from the ANNz2 package, trained on the Galaxy And Mass Assembly (GAMA) spectroscopy. Here we considerably improve over these previous redshift estimations by building a DL model, Hybrid-z, which combines four-band KiDS images with nine-band magnitudes from KiDS+VIKING. The Hybrid-z framework provides photo-$z$s for KiDS-Bright, with negligible mean residuals of O($10^{-4}$) and scatter at the level of $0.014(1+z)$ -- reduction by 20% over the previous nine-band derivations with ANNz2. We check our photo-$z$ model performance on test data drawn from GAMA, as well as from other KiDS-overlapping wide-angle spectroscopic surveys, namely SDSS, 2dFLenS, and 2dFGRS. We find stable behavior and consistent improvement over ANNz2 throughout. We finally apply Hybrid-z trained on GAMA to the entire KiDS-Bright DR4 sample of 1.2 million galaxies. For these final predictions, we design a method of smoothing the input redshift distribution of the training set, to avoid propagation of features present in GAMA, related to its small sky area and large-scale structure imprint in its fields. Our work paves the way towards the best-possible photo-$z$s achievable with machine learning for any galaxy type both for the final KiDS-Bright DR5 data and for future deeper imaging, such as from the Legacy Survey of Space and Time.

Figures

Figures reproduced from arXiv: 2501.01942 by the authors.

Figure 1
Figure 1. Inception module used in this study to estimate the [PITH_FULL_IMAGE:figures/full_fig_p005_1.png] view at source ↗
Figure 3
Figure 3. Performance of the Huber loss function during the train [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figure 2
Figure 2. Architecture of the Hybrid-z model, which em [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: Comparison of the true spectroscopic redshifts with the [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]
Figure 5
Figure 5. Figure 5: Photometric redshift errors of Hybrid-z as a function of four quantities, from left to right, (a) photometric redshift as derived [PITH_FULL_IMAGE:figures/full_fig_p009_5.png]
Figure 6
Figure 6. Figure 6: Comparison of photometric redshift error distributions [PITH_FULL_IMAGE:figures/full_fig_p010_6.png]
Figure 7
Figure 7. Figure 7: Photo-z distribution of the KiDS-DR4 Bright sample (r < 20 mag) as predicted by ANNz2 (blue, B21) compared with photo-z estimates from Hybrid-z (this work) trained on the spec-z distribution of (a) KiDS x GAMA bright sources (original training sample) (b) smoothed KiDS…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

19 extracted references · 8 canonical work pages

  1. [1]

    2020, ApJS, 249, 3, doi:10

    Ahumada, R., Allende Prieto, C., Almeida, A., et al. 2020, ApJS, 249, 3, doi:10. 3847/1538-4365/ab929e Ansari, Z., Agnello, A., & Gall, C. 2021, A&A, 650, A90, doi: 10.1051/ 0004-6361/202039675 Baum, W. A. 1957, AJ, 62, 6, doi: 10.1086/107433 Baum, W. A. 1962, in Problems of Extra-Galactic Research, ed. G. C. McVittie, V ol. 15, 390 Bellstedt, S., Driver,...

    doi:10.1086/107433 2020

  2. [3]

    2021, A&A, 653, A82, doi:

    1051/0004-6361/201731942 Bilicki, M., Dvornik, A., Hoekstra, H., et al. 2021, A&A, 653, A82, doi:

    work page 2021

  3. [4]

    2016, MNRAS, 462, 4240, doi:

    1051/0004-6361/202140352 Blake, C., Amon, A., Childress, M., et al. 2016, MNRAS, 462, 4240, doi:

    work page 2016

  4. [5]

    G., Sun, Y ., Davey, N., et al

    1093/mnras/stw1990 Bonfield, D. G., Sun, Y ., Davey, N., et al. 2010, Monthly Notices of the Royal As- tronomical Society, 405, 987, doi: 10.1111/j.1365-2966.2010.16544.x Capaccioli, M., & Schipani, P. 2011, The Messenger, 146, 2 Carliles, S., Budavári, T., Heinis, S., Priebe, C., & Szalay, A. S. 2010, ApJ, 712, 511, doi: 10.1088/0004-637X/712/1/511 Cavuo...

    arXiv 2010

  5. [6]

    A., & Lahav, O

    1046/j.1365-8711.2001.04902.x Collister, A. A., & Lahav, O. 2004, PASP, 116, 345, doi:10.1086/383254 Connolly, A. J., Csabai, I., Szalay, A. S., et al. 1995, AJ, 110, 2655, doi:

    arXiv 2001

  6. [7]

    1086/117720 Cunha, P. A. C., & Humphrey, A. 2022, A&A, 666, A87, doi: 10.1051/ 0004-6361/202243135 de Jong, J. T. A., Verdoes Kleijn, G. A., Kuijken, K. H., & Valentijn, E. A. 2013, Experimental Astronomy, 35, 25, doi: 10.1007/s10686-012-9306-1 de Jong, J. T. A., Verdoes Kleijn, Gijs A., Boxhoorn, Danny R., et al. 2015, A&A, 582, A62, doi: 10.1051/0004-63...

    arXiv doi:10.48550/arxiv.1611.00036 2022

  7. [8]

    A., McNaught-Roberts, T., et al

    1051/0004-6361/202245158 Eardley, E., Peacock, J. A., McNaught-Roberts, T., et al. 2015, MNRAS, 448, 3665, doi: 10.1093/mnras/stv237 Edge, A., Sutherland, W., Kuijken, K., et al. 2013, The Messenger, 154, 32 Gholamalinezhad, H., & Khosravi, H. 2020, arXiv e-prints, arXiv:2009.07485, doi: 10.48550/arXiv.2009.07485 Goodfellow, I., Bengio, Y ., & Courville, ...

    arXiv doi:10.48550/arxiv.2009.07485 2015

  8. [9]

    A., & Cai, Y .-C

    1051/0004-6361/202449929 Hang, Q., Alam, S., Peacock, J. A., & Cai, Y .-C. 2021, MNRAS, 501, 1481, doi: 10.1093/mnras/staa3738 Harris, C. R., Millman, K. J., van der Walt, S. J., et al. 2020, Nature, 585, 357, doi: 10.1038/s41586-020-2649-2 He, K., Zhang, X., Ren, S., & Sun, J. 2015, arXiv e-prints, arXiv:1512.03385, doi: 10.48550/arXiv.1512.03385 Henghes...

    arXiv doi:10.48550/arxiv.1512.03385 2021

Show all 19 references
  1. [10]

    3847/0067-0049/225/1/5 Bilicki, M., Hoekstra, H., Brown, M. J. I., et al. 2018, A&A, 616, A69, doi:

    2018

  2. [11]

    L., Wright, A

    1051/0004-6361/201014885 Hildebrandt, H., van den Busch, J. L., Wright, A. H., et al. 2021, A&A, 647, A124, doi: 10.1051/0004-6361/202039018 Hoyle, B. 2016, Astronomy and Computing, 16, 34, doi: https://doi.org/ 10.1016/j.ascom.2016.03.006 Huber, P. J. 1964, The Annals of Math...

    2021

  3. [12]

    1998, Proc

    1051/0004-6361/201834918 LeCun, Y ., Bottou, L., Bengio, Y ., & Haffner, P. 1998, Proc. IEEE, 86,

    1998

  4. [13]

    A., Allison, J., et al

    1093/mnras/stv1436 LSST Science Collaboration, Abell, P. A., Allison, J., et al. 2009, arXiv e-prints, arXiv:0912.0201, doi: 10.48550/arXiv.0912.0201 McCulloch, W.S., P. 1943, Bulletin of Mathematical Biophysics 5, 115–133 McFarland, J. P., Verdoes-Kleijn, G., Sikkema, G., et ...

    2009 doi arXiv

  5. [15]

    B., & Lahav, O

    1093/mnras/stw1281 Sadeh, I., Abdalla, F. B., & Lahav, O. 2016, PASP, 128, 104502 Schlafly, E. F., & Finkbeiner, D. P. 2011, The Astrophysical Journal, 737, 103, doi: 10.1088/0004-637X/737/2/103 Schlegel, D. J., Finkbeiner, D. P., & Davis, M. 1998, ApJ, 500, 525, doi:

    2016 doi

  6. [16]

    H., Cañameras, R., et al

    1086/305772 Schuldt, S., Suyu, S. H., Cañameras, R., et al. 2021, A&A, 651, A55, doi:

    2021

  7. [17]

    1051/0004-6361/202039945 Shorten, C., & Khoshgoftaar, T. M. 2019, Journal of Big Data, 6,

    2019

  8. [18]

    R., et al

    https: //api.semanticscholar.org/CorpusID:195811894 Strateva, I., Ivezi´c, Ž., Knapp, G. R., et al. 2001, AJ, 122, 1861, doi: 10.1086/ 323301 Szegedy, C., Liu, W., Jia, Y ., et al. 2015, Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 1 Ta...

    2001 doi

  9. [19]

    E., et al

    1051/0004-6361/202039293 Virtanen, P., Gommers, R., Oliphant, T. E., et al. 2020, Nature Methods, 17, 261, doi: 10.1038/s41592-019-0686-2 Wadadekar, Y . 2004, Publications of the Astronomical Society of the Pacific, 117, 79, doi: 10.1086/427710 Way, M. J., & Klose, C. D. 2012,...

    2020

  10. [2278]

    2021, Monthly Notices of the Royal Astronomi- cal Society, 509, 2289, doi: 10.1093/mnras/stab3165 Li, R., Napolitano, N

    https://api.semanticscholar.org/CorpusID:14542261 Li, C., Zhang, Y ., Cui, C., et al. 2021, Monthly Notices of the Royal Astronomi- cal Society, 509, 2289, doi: 10.1093/mnras/stab3165 Li, R., Napolitano, N. R., Feng, H., et al. 2022, A&A, 666, A85, doi: 10.1051/ 0004-6361/2022...

    2021 doi

  11. [2825]

    1996, in Neural Networks

    http://jmlr.org/papers/v12/ pedregosa11a.html Prechelt, L. 1996, in Neural Networks. https://api.semanticscholar. org/CorpusID:14049040 Roster, W., Salvato, M., Krippendorf, S., et al. 2024, A&A, 692, A260, doi:10. 1051/0004-6361/202452361 Rozo, E., Ryko ff, E. S., Abate, A., ...

    1996

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.