Pith. sign in

REVIEW 3 major objections 5 minor 90 references

An Unsupervised Machine Learning Approach to Identify Spectral Energy Distribution Outliers: Application to the S-PLUS DR4 data

T0 review · 3 major / 5 minor · reviewed 2026-08-16 · deepseek-v4-flash

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

Pith's one-line read An unsupervised autoencoder plus t-SNE pipeline on S-PLUS photometry flags roughly 19,000 anomalous SEDs and isolates 69 carbon-rich candidates, along with distinct white-dwarf and active low-mass star populations.

desk verdict A useful, transparent candidate-selection pipeline for S-PLUS; the carbon-star candidates have external support, but the t-SNE perplexity tuning and the MG-augmented WD claim need harder stability numbers. read the letter →

arxiv 2504.18491 v1 pith:Q72HASI3 submitted 2025-04-25 astro-ph.SR astro-ph.GAastro-ph.IM

F. Quispe-Huaynasi , F. Roig , N. Holanda , V. Loaiza-Tacuri , Romualdo Eleutério , C. B. Pereira , S. Daflon , V. M. Placco
show 7 more authors
R. Lopes de Oliveira F. Sestito P. K. Humire M. Borges Fernandes A. Kanaan T. Ribeiro W. Schoenell
This is my paper · ORCID
classification astro-ph.SRastro-ph.GAastro-ph.IM
keywords autoencodert-SNES-PLUSspectralenergydistributionanomalydetectioncarbonstarswhitedwarfslow-mass
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

This paper tries to establish that a purely unsupervised pipeline, an autoencoder followed by t-SNE, can use 12-band photometry from the S-PLUS survey to produce a short list of genuinely peculiar stars worth spectroscopic follow-up. The authors train an autoencoder on about 1.5 million stars they believe are normal, then flag the roughly 19,000 sources whose spectral energy distributions the network cannot reconstruct beyond a $3\sigma$ error threshold. Running t-SNE on the 66 colors of the anomalous sample, they identify 69 carbon-rich candidates (likely CH or carbon-enhanced metal-poor stars, four possibly dwarf carbon stars), separate DA, DB, and WD-plus-main-sequence white dwarf groups, and isolate very active low-mass stars using X-ray data. If correct, the method turns a broad-band plus narrow-band photometric survey into a discovery engine for chemically peculiar stars, compact remnants, and active stars, independent of spectroscopy.

What carries the argument

The machinery is a two-stage unsupervised pipeline. First, a vanilla autoencoder with a seven-dimensional latent code is trained on about 1.5 million stars that pass strict astrometric and photometric quality cuts; its mean squared reconstruction error on the 12 scaled magnitudes defines normality, and a $3\sigma$ threshold selects roughly 19,000 anomalies. Second, t-SNE is run on the 66 color differences among the 12 S-PLUS filters, with perplexity set to 40, to lay the anomalies out on a 2D map where known populations form dense islands; for white dwarfs, an absolute magnitude from an external astrometric catalog is added as an extra feature to separate them from hot subdwarfs.

What would settle it

Spectroscopically observe the 69 carbon-rich candidates: if most turn out to lack enhanced carbon features such as CH or C2 bands, and if their space velocities place them in the thin disk rather than the halo or thick disk, then the claimed CH/CEMP identification collapses.

Watch

Extended reading notes

Core claim

The central discovery is that SED outliers selected by reconstruction error are not a random collection: in the t-SNE projection they form coherent islands that match established stellar classes. The autoencoder's $3\sigma$ anomalies, when projected through t-SNE on 66 color indices, cluster at positions occupied by known carbon stars, DA and DB white dwarfs, WD-plus-main-sequence binaries, and X-ray-active low-mass stars. The paper's key population claim is that 69 of the carbon candidates belong, by kinematics and space position, to the CH/CEMP family, with four being dwarf carbon stars. It also reports that binary systems are abundant among anomalies but show no clean relation between t-SNE overdensities and orbital parameters.

Load-bearing premise

The result depends on the quality cuts leaving a training set of about 1.5 million genuinely normal, single, non-variable stars; if unresolved binaries, variables, or misclassified extragalactic sources survive, the autoencoder learns a biased idea of normal and every anomaly label inherits that bias.

Editorial extensions

If this is right

  • The 69 carbon-rich candidates form a concrete target list for medium- and high-resolution spectroscopy; confirmed CH or CEMP stars would expand the known chemically peculiar population in the southern sky.
  • The clean separation of DA, DB, and WD+MS groups implies that S-PLUS colors, plus a distance, can pre-select white-dwarf subtypes for follow-up without needing spectra.
  • The X-ray-active low-mass stars occupy a distinct t-SNE region, so the pipeline can flag active stars from photometry alone, even before epoch photometry or activity-line indices are examined.
  • Population sizes and the anomaly sample itself depend on the $3\sigma$ threshold and the t-SNE perplexity, so the outputs are candidate lists rather than complete censuses.

Reading between the lines

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

  • Because the t-SNE perplexity was chosen to maximize the carbon-star candidate count, other cluster boundaries on the same map may be partly optimized for that population; re-running t-SNE across several perplexities and asking which clusters persist would test whether the white-dwarf and active-star groupings are stable.
  • The null result on binary orbital properties suggests SED shape is dominated by photospheric parameters rather than orbital architecture; a forward model that generates synthetic SEDs from binary parameters could quantify how many SB1 systems this pipeline is intrinsically blind to.
  • The same autoencoder-plus-t-SNE recipe should transfer directly to surveys sharing the S-PLUS filter set, so a cross-survey run would reveal whether the identified populations are survey-independent or artifacts of S-PLUS depth and footprint.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 5 minor

Summary. The manuscript presents an unsupervised machine-learning pipeline for identifying spectral energy distribution (SED) outliers in S-PLUS DR4 photometry. An autoencoder is trained on roughly 1.5 million stars selected by a series of quality cuts, and about 19,000 objects whose SEDs are poorly reconstructed are flagged as anomalous. t-SNE is then applied to 66 S-PLUS colors for the anomalous sample, and clusters are interpreted using cross-matches with Gaia, SIMBAD, eROSITA, APOGEE, and external carbon-star catalogs. The authors report 96 carbon-star candidates (69 newly identified, 4 likely dwarf carbon stars), three white-dwarf subgroups (DA, DB, WD+MS binaries), active low-mass stars detected in X-rays, and a large number of binary systems. The results are framed as candidate lists for spectroscopic follow-up, and a machine-readable candidate table is provided.

Significance. If the method is robust, the paper offers a practical demonstration that unsupervised anomaly detection plus nonlinear dimensionality reduction on narrow-band S-PLUS photometry can produce useful spectroscopic follow-up targets. The manuscript is grounded in real survey data and makes appropriate use of external cross-matches: the Lucey et al. (2023) match strengthens the CEMP association for known objects, the eROSITA data support the active-star interpretation, and the comparison with Li et al. (2024) helps place the carbon-star candidates in context. The code and exact hyperparameters are described in enough detail to reproduce the pipeline. The principal weakness is that the headline candidate counts are partly determined by a t-SNE hyperparameter chosen to maximize the carbon-star count, so the quantitative claims need a robustness analysis before they can be taken at face value.

major comments (3)
  1. [§4.1 and §5.1] The choice of t-SNE perplexity is made after inspecting the outcome: the text states that perplexity 40 'was mostly motivated because it yielded the highest number of carbon stars candidates.' The subsequently reported number of 96 C-star candidates (69 newly identified) is therefore a selected maximum over the tested configurations, not an unbiased estimate. Because t-SNE is stochastic and its geometry depends on perplexity, this post-hoc selection also affects the white-dwarf and M-star clusters defined on the same map. Please provide a quantitative stability analysis: for each tested perplexity value (25, 30, 35, 40, 45) and, ideally, for several random seeds, report the total number of carbon-star candidates, the overlap or Jaccard index between the perplexity-40 list and the other lists, and the membership stability of the carbon-star and white-dwarf regions. The statement that variations are 'subtle' should be supported by these numbers, not only by visual inspection.
  2. [§5.2] The separation of white dwarfs from hot subdwarfs is achieved only after adding the absolute magnitude MG as an additional input feature to t-SNE. Since MG is a luminosity-dependent quantity, the resulting separation is not purely based on SED colors, and the manuscript should explicitly state that the WD sub-population segregation is partly driven by absolute magnitude. In addition, the assignment of overdensities 1, 2, and 3 to WD+MS binaries, DB/DC stars, and DA stars rests on visual inspection of SIMBAD spectral types. Please add a quantitative validation, for example a contingency table, purity/completeness of each region with respect to SIMBAD classes, or a significance test of the spatial segregation. Without such validation, the claim that t-SNE 'reliably' segregates WD sub-populations is not fully supported.
  3. [§3.2 and §6] The training set is designed to contain 'well-behaved' stars, but no diagnostic is shown that the autoencoder reconstruction errors on the training set are approximately Gaussian, which is the basis for the 3σ anomaly threshold. Since all subsequent populations are derived from this threshold, even a small fraction of contaminants in the training set could propagate into the outlier sample and into every population label. Please show the distribution of reconstruction errors and report the sensitivity of the anomalous sample size, and of the carbon-star and white-dwarf candidate lists, to the threshold (for example 2.5σ and 3.5σ). The discussion already acknowledges threshold dependence qualitatively, but the paper would be much stronger with a quantitative test.
minor comments (5)
  1. [Abstract and §5.1] The abstract reports 69 carbon-rich star candidates, while §5.1 says that 96 C-star candidates were identified, 27 of them previously reported. Please clarify explicitly that 69 is the number of newly identified candidates, and state whether the 73-object cross-match with Lucey et al. (2023) includes previously known objects.
  2. [§4.1] The phrase 'the number of iteration hyperparameters' should read 'the number of iterations hyperparameter'.
  3. [§5.3] 'Counter parts' should be 'counterparts' in the last paragraph of Section 5.3.
  4. [Figure 5 caption] The sentence 'The ChemPec* label encompass a variety...' should be 'The ChemPec* label encompasses a variety...'.
  5. [§3.2] The claim that photometric errors are 'much smaller' than the reconstruction-error threshold is asserted without numbers. Please quote typical reconstruction errors and typical photometric errors to support this statement.

Circularity Check

1 steps flagged · score 4.0 of 10

Carbon-star candidate count is a selected maximum: t-SNE perplexity was tuned to maximize that count, so the headline number is partly self-selected; population identification itself is externally anchored.

  1. fitted input called prediction [Section 4.1 (t-SNE applied to S-PLUS colors), with the resulting count reported in Section 5.1 and abstract]
    "Choosing a perplexity value of 40, in particular, was mostly motivated because it yielded the highest number of carbon stars candidates."

    The t-SNE hyperparameter was chosen by maximizing the number of carbon-star candidates, and the same quantity (96 candidates, 69 reported as CH/CEMP) is then presented as the paper's discovery. At the chosen perplexity, the candidate count is by construction at least as large as at the other tested perplexities (25, 35, 45), so the reported number is a selected maximum rather than an unbiased estimate. The paper states that variations across perplexity are 'subtle' but gives no quantitative candidate-list stability or seed-dependence analysis. This does not invalidate the external SIMBAD/Lucey et al. corroboration of individual candidates, but the headline count is partially an artifact of the tuning procedure.

full rationale

The paper is largely self-contained: the autoencoder is trained on a quality-cut sample, the anomaly threshold is a stated 3-sigma criterion, and the stellar-population labels (carbon stars, white dwarfs, active low-mass stars, binaries) are anchored to external catalogs such as SIMBAD, APOGEE, Gaia non-single-star tables, and eROSITA X-ray data. The main circularity concern is confined to the carbon-star candidate count: perplexity 40 was explicitly chosen because it yielded the highest number of carbon-star candidates, and that count is then reported as a result. This is a form of selection on the dependent variable rather than a hidden fit, and the paper does disclose the choice; however, without a quantitative stability analysis the headline 96/69 count is partly a selected maximum. The white-dwarf and active-star identifications are validated by external labels and do not reduce to the method's inputs. Self-citations in the paper (e.g., S-PLUS and classification catalogs) are used as data references, not as load-bearing uniqueness arguments, so they do not increase the circularity score beyond the one identified step.

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

The central results depend on a chain of modeling choices: the definition of 'normal' stars, the anomaly threshold, and the t-SNE hyperparameters. The most consequential free parameter is the perplexity, which was tuned to maximize the carbon-star count, so the reported number of candidates is partly an artifact of that choice. No new physical entities are introduced.

free parameters (4)
  • Reconstruction error threshold = 3 sigma above the mean training reconstruction error
    Chosen by hand as a conservative trade-off. The size of the anomalous sample (~19,000) changes with this threshold, as the paper notes in Section 6.
  • t-SNE perplexity = 40
    Selected after testing 25, 30, 35, 40, and 45 because it yielded the highest number of carbon star candidates (Section 4.1). This directly influences the reported count of 69 candidates.
  • MG > 5.0 mag criterion for dwarf carbon stars = 5.0 mag
    Adopted in Section 5.1 to classify four objects as likely carbon dwarfs. It is a literature-motivated boundary but still a chosen cutoff.
  • Autoencoder architecture hyperparameters = Two hidden layers of 8 nodes, latent dimension 7, Adam with learning rate 0.001
    Tuned automatically with KerasTuner RandomSearch, but no independent validation set is reported. The central anomaly detection claim is not highly sensitive to small changes, but these are selected values.
assumptions (4)
  • domain assumption The quality cuts in Section 3.1 select only genuine single, non-variable stars with clean astrometry and no extragalactic contamination.
    The autoencoder is trained on the resulting ~1.5 million sources as the definition of 'normal.' If binaries, variables, or misclassified galaxies remain, the learned reconstruction error distribution is biased.
  • domain assumption The autoencoder trained on stars brighter than magnitude 19 generalizes to fainter sources in the survey.
    Section 6 acknowledges that fainter sources may be flagged as anomalous simply because they are outside the training magnitude range, weakening the anomaly interpretation.
  • standard math t-SNE with perplexity 40 and 2000 iterations preserves the local neighborhood structure of the 66-color space.
    This is the standard t-SNE assumption, but the paper verifies stability only qualitatively ('overall structure and densities remain mostly unchanged').
  • domain assumption SIMBAD classifications used to label clusters are reliable and complete enough for the populations under study.
    The paper cross-matches with SIMBAD and uses the primary type labels, but SIMBAD is heterogeneous and incomplete, and the paper excludes generic 'Star' labels.

how reviews work

0 comments
Cite this review

Pith. "Pith review of An Unsupervised Machine Learning Approach to Identify Spectral Energy Distribution Outliers: Application to the S-PLUS DR4 data." pith.science (2026). https://pith.science/paper/Q72HASI3

@misc{pith2026250418491,
  author       = {Pith},
  title        = {Pith review of: An Unsupervised Machine Learning Approach to Identify Spectral Energy Distribution Outliers: Application to the S-PLUS DR4 data},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/Q72HASI3}},
  note         = {Machine review of arXiv:2504.18491}
}
abstract

Identification of specific stellar populations using photometry for spectroscopic follow-up is a first step to confirm and better understand their nature. In this context, we present an unsupervised machine learning approach to identify candidates for spectroscopic follow-up using data from the Southern Photometric Local Universe Survey (S-PLUS). First, using an anomaly detection technique based on an autoencoder model, we select a large sample of objects ($\sim 19,000$) whose Spectral Energy Distribution (SED) is not well reconstructed by the model after training it on a well-behaved star sample. Then, we apply the t-distributed Stochastic Neighbor Embedding (t-SNE) algorithm to the 66 color measurements from S-PLUS, complemented by information from the SIMBAD database, to identify stellar populations. Our analysis reveals 69 carbon-rich star candidates that, based on their spatial and kinematic characteristics, may belong to the CH or Carbon-Enhanced Metal-Poor (CEMP) categories. Among these chemically peculiar candidates, we identify four as likely carbon dwarf stars. We show that it is feasible to identify three primary white dwarf (WD) populations: WDs with hydrogen-dominated atmospheres (DA), WDs with neutral helium-dominated atmospheres (DB), and the WDs main sequence binaries (WD + MS). Furthermore, by using eROSITA X-ray data, we also highlight the identification of candidates for very active low-mass stars. Finally, we identified a large number of binary systems using the autoencoder model, but did not observe a clear association between the overdensities in the t-SNE map and their orbital properties.

Figures

Figures reproduced from arXiv: 2504.18491 by the authors.

Figure 1
Figure 1. (a) Color-magnitude diagram for the training sample (∼ 1.5 million stars). (b) Kiel diagram for objects in common between the training sample and SIMBAD, colored by metallicity (∼ 20, 000 stars). The stellar parameters are those given in SIMBAD, and come from a diversity of independent works. Input Layer Encoder Layer 1 Encoder Layer 2 Code Decoder Layer 2 Decoder Layer 1 Output Layer 12 features 8 nodes 8 nodes 7 n… view at source ↗
Figure 2
Figure 2. Schematic representation of the autoencoder architecture after hyperparameter tuning. 4000 5000 6000 7000 8000 9000 Wavelength ˚A 0.5 0.6 0.7 0.8 0.9 Scaled Magnitute Normal Source Input Output 4000 5000 6000 7000 8000 9000 Wavelength ˚A 0.45 0.50 0.55 0.60 0.65 Scaled Magnitude Anomalous Source Input Output [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Comparisons between input spectral energy distribution (SED), in blue, and the reconstructed SED, in orange, for a normal source (left panel) and for an anomalous source (right panel). at xi . The value of σi for each xi is determined inter- nally by t-SNE through a binary search algorithm based [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (9 more)
Figure 4
Figure 4. Figure 4: t-SNE 2D projections. (a) Colored by Gaia ruwe parameter. (b) Colored by Gaia absolute magnitude. The maps contain all the sources identified using the autoencoder. Except for the perplexity and the number of iterations, which were set to 40 and 2000, respectively, all…
Figure 5
Figure 5. Figure 5: t-SNE 2D maps: In all panels, the background in gray represents sources identified by the autoencoder. The colored points indicate sources that overlap with sources classified in the SIMBAD database. The ChemPec* label encompass a variety of chemically peculiar objects…
Figure 6
Figure 6. Figure 6: Color-magnitude diagram. In all panels, the background in gray represents sources identified by the autoencoder. The colored points indicate sources that overlap with sources classified in the SIMBAD database. The ChemPec* label encompass a variety of chemically peculi…
Figure 7
Figure 7. Figure 7: Spectral energy distribution of the populations described in this work specific spectral lines such as Ca H + K, Hδ, Hα, and the Ca triplet, which are indicators of stellar activity (Cincunegui et al. 2007). 5.4. Binary systems Because quality cuts, such as astrometric…
Figure 8
Figure 8. Figure 8: Comparison of C star candidates (red circles) with the C stars analyzed by Li et al. (2024). Panel (a) presents a color-magnitude diagram with different types of C stars color-coded: green for C-N, orange for CH, and blue for C-R. A dashed line at MG = 5.0 mag marks th…
Figure 9
Figure 9. Figure 9: Zooming in on the region of carbon-rich stars in the t-SNE map. The carbon-rich stars identified in this study are shown in gray. The red, green, and purple points indicate stars classified as carbon, peculiar, and RGB in SIMBAD, respectively. normal in this work, whil…
Figure 10
Figure 10. Figure 10: White dwarf populations. The left panel shows the t-SNE map, illustrating the separation between WDs (blue dots) and hot subdwarf stars (orange dots) after using the MG as an additional feature for the t-SNE input. The regions 1, 2, and 3 in this panel are over-densit…
Figure 12
Figure 12. Figure 12: t-SNE 2D map constructed for a set that in￾cludes all the anomalous stars detected by the autoencoder (blue dots) and a subset of the normal stars used to train the autoencoder (green dots). The location of the anoma￾lous stars identified as spectroscopic binaries in …
Figure 11
Figure 11. Figure 11: The upper panel shows a zoomed-in view of the low-mass star region in the t-SNE map, with sources con￾taining X-ray information highlighted in orange. The bot￾tom panel displays the X-ray main-sequence diagram, where objects with X-ray data are positioned in the expec…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

90 extracted references · 10 canonical work pages

  1. [1]

    , " * write output.state after.block = add.period write newline

    ENTRY address archivePrefix author booktitle chapter doi edition editor eprint howpublished institution journal key month number organization pages publisher school series title misctitle type volume year version url label extra.label sort.label short.list INTEGERS output.state before.all mid.sentence after.sentence after.block FUNCTION init.state.consts ...

  2. [2]

    write newline

    " write newline "" before.all 'output.state := FUNCTION format.url url empty "" new.block "" url * "" * if FUNCTION format.eprint eprint empty "" archivePrefix empty "" archivePrefix "arXiv" = new.block " " eprint * " " * new.block " " eprint * " " * if if if FUNCTION format.doi doi empty "" " " doi * " " * if FUNCTION format.pid doi empty eprint empty ur...

  3. [3]

    ]׿=| q?0 ;u8lT Q ͙ GfW* 'k) =QcZR F Q a hE: TtPN :.9u ੇ< j˄AJx Ô&u).4Z'u S'f j Z>lC ӱ¬ 5a XAlA'2f|7EjĔǤ 3P G

    thebibliography [1] 20pt to REFERENCES 6pt =0pt -12pt 10pt plus 3pt =0pt =0pt =1pt plus 1pt =0pt =0pt -12pt =13pt plus 1pt =20pt =13pt plus 1pt \@M =10000 =-1.0em =0pt =0pt 0pt =0pt =1.0em @enumiv\@empty 10000 10000 `\.\@m \@noitemerr \@latex@warning Empty `thebibliography' environment \@ifnextchar \@reference \@latexerr Missing key on reference command E...

    arXiv 2021

  4. [4]

    2022, , 664, A45, 10.1051/0004-6361/202243595

    Abia , C., de Laverny , P., Romero-G \'o mez , M., & Figueras , F. 2022, , 664, A45, 10.1051/0004-6361/202243595

    doi:10.1051/0004-6361/202243595 2022

  5. [5]

    2000, , 536, 438, 10.1086/308932

    Abia , C., & Isern , J. 2000, , 536, 438, 10.1086/308932

    doi:10.1086/308932 2000

  6. [6]

    2016, , 595, A129, 10.1051/0004-6361/201628789

    Allende Prieto , C. 2016, , 595, A129, 10.1051/0004-6361/201628789

    doi:10.1051/0004-6361/201628789 2016

  7. [7]

    X., et al

    Anders , F., Chiappini , C., Santiago , B. X., et al. 2018, , 619, A125, 10.1051/0004-6361/201833099

    doi:10.1051/0004-6361/201833099 2018

  8. [8]

    Bailer-Jones , C. A. L., Rybizki , J., Fouesneau , M., Demleitner , M., & Andrae , R. 2021, VizieR Online Data Catalog, I/352

    2021

Show all 90 references
  1. [9]

    2023, Autoencoders (Cham: Springer International Publishing), 353--374, 10.1007/978-3-031-24628-9_16

    Bank, D., Koenigstein, N., & Giryes, R. 2023, Autoencoders (Cham: Springer International Publishing), 353--374, 10.1007/978-3-031-24628-9_16

    2023 doi

  2. [10]

    C., & Christlieb , N

    Beers , T. C., & Christlieb , N. 2005, , 43, 531, 10.1146/annurev.astro.42.053102.134057

    2005

  3. [11]

    2014, arXiv e-prints, arXiv:1403.5237, 10.48550/arXiv.1403.5237

    Benitez , N., Dupke , R., Moles , M., et al. 2014, arXiv e-prints, arXiv:1403.5237, 10.48550/arXiv.1403.5237

    2014 doi arXiv

  4. [12]

    S., Li, Y., & Xu, Y

    Berahmand, K., Daneshfar, F., Salehi, E. S., Li, Y., & Xu, Y. 2024, Artificial Intelligence Review, 57, 28, 10.1007/s10462-023-10662-6

    2024 doi

  5. [13]

    2024, arXiv e-prints, arXiv:2409.03941, 10.48550/arXiv.2409.03941

    Blouin , S. 2024, arXiv e-prints, arXiv:2409.03941, 10.48550/arXiv.2409.03941

    2024 doi arXiv

  6. [14]

    H., MacAlpine , G., et al

    Bothun , G., Elias , J. H., MacAlpine , G., et al. 1991, , 101, 2220, 10.1086/115844

    1991 doi

  7. [15]

    J., Moles , M., Crist \'o bal-Hornillos , D., et al

    Cenarro , A. J., Moles , M., Crist \'o bal-Hornillos , D., et al. 2019, , 622, A176, 10.1051/0004-6361/201833036

    2019 doi

  8. [16]

    C., Magnier , E

    Chambers , K. C., Magnier , E. A., Metcalfe , N., et al. 2016, arXiv e-prints, arXiv:1612.05560, 10.48550/arXiv.1612.05560

    2016 doi arXiv

  9. [17]

    2015, Keras, https://keras.io

    Chollet, F., et al. 2015, Keras, https://keras.io

    2015

  10. [18]

    J., Wisotzki , L., & Reimers , D

    Christlieb , N., Green , P. J., Wisotzki , L., & Reimers , D. 2001, , 375, 366, 10.1051/0004-6361:20010814

    2001 doi

  11. [19]

    F., & Mauas , P

    Cincunegui , C., D \' az , R. F., & Mauas , P. J. D. 2007, , 469, 309, 10.1051/0004-6361:20066503

    2007 doi

  12. [20]

    2012, Res

    Cui , X.-Q., Zhao , Y.-H., Chu , Y.-Q., et al. 2012, Res. Astron. Astrophys., 12, 1197, 10.1088/1674-4527/12/9/003

    2012 doi

  13. [21]

    S., Bessell , M

    Da Costa , G. S., Bessell , M. S., Mackey , A. D., et al. 2019, , 489, 5900, 10.1093/mnras/stz2550

    2019 doi

  14. [22]

    M., Freeman , K

    De Silva , G. M., Freeman , K. C., Bland-Hawthorn , J., et al. 2015, , 449, 2604, 10.1093/mnras/stv327

    2015 doi

  15. [23]

    Dominy , J. F. 1984, , 55, 27, 10.1086/190946

    1984 doi

  16. [24]

    W., Riello , M., De Angeli , F., et al

    Evans , D. W., Riello , M., De Angeli , F., et al. 2018, , 616, A4, 10.1051/0004-6361/201832756

    2018 doi

  17. [25]

    E., Guti \'e rrez-Soto , L

    Ferreira Lopes , C. E., Guti \'e rrez-Soto , L. A., Ferreira Alberice , V. S., et al. 2024, arXiv e-prints, arXiv:2411.18748, 10.48550/arXiv.2411.18748

    2024 doi arXiv

  18. [26]

    2000, , 120, 2615, 10.1086/316824

    Finlator , K., Ivezi \'c , Z ., Fan , X., et al. 2000, , 120, 2615, 10.1086/316824

    2000 doi

  19. [27]

    2018, , 616, A10, 10.1051/0004-6361/201832843

    Gaia Collaboration , Babusiaux , C., van Leeuwen , F., et al. 2018, , 616, A10, 10.1051/0004-6361/201832843

    2018 doi

  20. [28]

    Gaia Collaboration , Vallenari , A., Brown , A. G. A., et al. 2023 a , , 674, A1, 10.1051/0004-6361/202243940

    2023 doi

  21. [29]

    2023 b , , 674, A34, 10.1051/0004-6361/202243782

    Gaia Collaboration , Arenou , F., Babusiaux , C., et al. 2023 b , , 674, A34, 10.1051/0004-6361/202243782

    2023 doi

  22. [30]

    Green , G. 2018, J. Open Source Soft., 3, 695, 10.21105/joss.00695

    2018 doi

  23. [31]

    2004, , 12, 71, 10.1007/s00159-004-0023-2

    G \"u del , M. 2004, , 12, 71, 10.1007/s00159-004-0023-2

    2004 doi

  24. [32]

    R., Majewski , S

    Hayes , C. R., Majewski , S. R., Shetrone , M., et al. 2018, , 852, 49, 10.3847/1538-4357/aa9cec

    2018 doi

  25. [33]

    2003, , 586, 195, 10.1086/367536

    Helmi , A., Ivezi \'c , Z ., Prada , F., et al. 2003, , 586, 195, 10.1086/367536

    2003 doi

  26. [34]

    R., Almeida-Fernandes , F., Oliveira Schwarz , G

    Herpich , F. R., Almeida-Fernandes , F., Oliveira Schwarz , G. B., et al. 2024, , 689, A249, 10.1051/0004-6361/202449725

    2024 doi

  27. [35]

    2005, , 43, 435, 10.1146/annurev.astro.43.072103.150600

    Herwig , F. 2005, , 43, 435, 10.1146/annurev.astro.43.072103.150600

    2005 arXiv

  28. [36]

    E., & Roweis, S

    Hinton, G. E., & Roweis, S. 2002, in Advances in Neural Information Processing Systems, ed. S. Becker, S. Thrun, & K. Obermayer, Vol. 15 (MIT Press). https://proceedings.neurips.cc/paper_files/paper/2002/file/6150ccc6069bea6b5716254057a194ef-Paper.pdf

    2002

  29. [38]

    E., & Salakhutdinov, R

    Hinton, G. E., & Salakhutdinov, R. R. 2006, Science, 313, 504, 10.1126/science.1127647

    2006 doi

  30. [39]

    C., Xiao , K., et al

    Huang , Y., Beers , T. C., Xiao , K., et al. 2024, arXiv e-prints, arXiv:2408.02171, 10.48550/arXiv.2408.02171

    2024 doi arXiv

  31. [40]

    Jura , M., & Young , E. D. 2014, Annual Review of Earth and Planetary Sciences, 42, 45, 10.1146/annurev-earth-060313-054740

    2014 doi

  32. [41]

    L., Hawkins , K., Rogers , L

    Kao , M. L., Hawkins , K., Rogers , L. K., et al. 2024, , 970, 181, 10.3847/1538-4357/ad5d6e

    2024 doi

  33. [42]

    P., & Ba , J

    Kingma , D. P., & Ba , J. 2014, arXiv e-prints, arXiv:1412.6980. 1412.6980

    2014 arXiv

  34. [43]

    2001, , 371, 222, 10.1051/0004-6361:20010348

    Knapp , G., Pourbaix , D., & Jorissen , A. 2001, , 371, 222, 10.1051/0004-6361:20010348

    2001 doi

  35. [44]

    Laroche , A., & Speagle , J. S. 2025, , 979, 5, 10.3847/1538-4357/ad9607

    2025 doi

  36. [45]

    D., Newberg , J., Rosner , R., Richards , G

    Lenz , D. D., Newberg , J., Rosner , R., Richards , G. T., & Stoughton , C. 1998, , 119, 121, 10.1086/313155

    1998 doi

  37. [46]

    2024, , 271, 12, 10.3847/1538-4365/ad1881

    Li , L., Zhang , K., Cui , W., et al. 2024, , 271, 12, 10.3847/1538-4365/ad1881

    2024 doi

  38. [47]

    E., Ederoclite , A., et al

    L \'o pez-Sanjuan , C., Tremblay , P. E., Ederoclite , A., et al. 2022, , 658, A79, 10.1051/0004-6361/202141746

    2022 doi

  39. [48]

    2023, , 523, 4049, 10.1093/mnras/stad1675

    Lucey , M., Al Kharusi , N., Hawkins , K., et al. 2023, , 523, 4049, 10.1093/mnras/stad1675

    2023 doi

  40. [49]

    R., Schiavon , R

    Majewski , S. R., Schiavon , R. P., Frinchaboy , P. M., et al. 2017, , 154, 94, 10.3847/1538-3881/aa784d

    2017 doi

  41. [50]

    C., Fanson , J., Schiminovich , D., et al

    Martin , D. C., Fanson , J., Schiminovich , D., et al. 2005, , 619, L1, 10.1086/426387

    2005 doi

  42. [51]

    2006, , 450, 735, 10.1051/0004-6361:20054021

    Masana , E., Jordi , C., & Ribas , I. 2006, , 450, 735, 10.1051/0004-6361:20054021

    2006 doi

  43. [52]

    Mauron , N., Azzopardi , M., Gigoyan , K., & Kendall , T. R. 2004, , 418, 77, 10.1051/0004-6361:20034264

    2004 doi

  44. [53]

    2019, , 489, 241, 10.1093/mnras/stz1985

    Mendes de Oliveira , C., Ribeiro , T., Schoenell , W., et al. 2019, , 489, 241, 10.1093/mnras/stz1985

    2019 doi

  45. [54]

    2024, , 682, A34, 10.1051/0004-6361/202347165

    Merloni , A., Lamer , G., Liu , T., et al. 2024, , 682, A34, 10.1051/0004-6361/202347165

    2024 doi

  46. [55]

    D., & Swart, T

    Mienye, I. D., & Swart, T. G. 2025, Archives of Computational Methods in Engineering, 10.1007/s11831-025-10260-5

    2025 doi

  47. [56]

    2003, , 583, 451, 10.1086/345097

    Mohanty , S., & Basri , G. 2003, , 583, 451, 10.1086/345097

    2003 doi

  48. [57]

    Nakazono , L., Mendes de Oliveira , C., Hirata , N. S. T., et al. 2021, , 507, 5847, 10.1093/mnras/stab1835

    2021 doi

  49. [58]

    M., Ghahramani, Z., & Valera, I

    Nazábal, A., Olmos, P. M., Ghahramani, Z., & Valera, I. 2020, Pattern Recognition, 107, 107501, https://doi.org/10.1016/j.patcog.2020.107501

    2020

  50. [59]

    R., Irwin , J., Charbonneau , D., et al

    Newton , E. R., Irwin , J., Charbonneau , D., et al. 2017, , 834, 85, 10.3847/1538-4357/834/1/85

    2017 doi

  51. [60]

    2019, Keras Tuner , https://github.com/keras-team/keras-tuner

    O'Malley, T., Bursztein, E., Long, J., et al. 2019, Keras Tuner , https://github.com/keras-team/keras-tuner

    2019

  52. [61]

    2018, in 2018 International Conference on Intelligent Systems and Computer Vision (ISCV), 1--8, 10.1109/ISACV.2018.8354080

    Oussidi, A., & Elhassouny, A. 2018, in 2018 International Conference on Intelligent Systems and Computer Vision (ISCV), 1--8, 10.1109/ISACV.2018.8354080

    2018

  53. [62]

    Pang, G., Shen, C., Cao, L., & Hengel, A. V. D. 2021, ACM Comput. Surv., 54, 10.1145/3439950

    2021 doi

  54. [63]

    2011, Journal of Machine Learning Research, 12, 2825

    Pedregosa, F., Varoquaux, G., Gramfort, A., et al. 2011, Journal of Machine Learning Research, 12, 2825

    2011

  55. [64]

    Portillo , S. K. N., Parejko , J. K., Vergara , J. R., & Connolly , A. J. 2020, , 160, 45, 10.3847/1538-3881/ab9644

    2020 doi

  56. [65]

    J., Menten , K

    Reid , M. J., Menten , K. M., Brunthaler , A., et al. 2014, , 783, 130, 10.1088/0004-637X/783/2/130

    2014 doi

  57. [66]

    Rodriguez , A. C. 2024, , 136, 054201, 10.1088/1538-3873/ad357c

    2024 doi

  58. [67]

    E., Hinton , G

    Rumelhart , D. E., Hinton , G. E., & Williams , R. J. 1986, , 323, 533, 10.1038/323533a0

    1986 doi

  59. [68]

    M., Rix , H.-W., et al

    Rybizki , J., Green , G. M., Rix , H.-W., et al. 2022, , 510, 2597, 10.1093/mnras/stab3588

    2022 doi

  60. [69]

    2022, , 988, 1, 10.1016/j.physrep.2022.09.001

    Saumon , D., Blouin , S., & Tremblay , P.-E. 2022, , 988, 1, 10.1016/j.physrep.2022.09.001

    2022 doi

  61. [70]

    F., & Finkbeiner , D

    Schlafly , E. F., & Finkbeiner , D. P. 2011, , 737, 103, 10.1088/0004-637X/737/2/103

    2011 doi

  62. [71]

    2019, Bulletin de la Societe Royale des Sciences de Liege, 88, 174

    Sharma , K., Kembhavi , A., Kembhavi , A., Sivarani , T., & Abraham , S. 2019, Bulletin de la Societe Royale des Sciences de Liege, 88, 174

    2019

  63. [72]

    G., Mironov , A

    Sichevskij , S. G., Mironov , A. V., & Malkov , O. Y. 2014, Astrophysical Bulletin, 69, 160, 10.1134/S1990341314020035

    2014 doi

  64. [73]

    F., Cutri , R

    Skrutskie , M. F., Cutri , R. M., Stiening , R., et al. 2006, , 131, 1163, 10.1086/498708

    2006 doi

  65. [74]

    Souto , D., Cunha , K., & Smith , V. V. 2021, , 917, 11, 10.3847/1538-4357/abfdb5

    2021 doi

  66. [75]

    V., et al

    Souto , D., Cunha , K., Smith , V. V., et al. 2022, , 927, 123, 10.3847/1538-4357/ac4891

    2022 doi

  67. [76]

    2017, , 471, 2587, 10.1093/mnras/stx1068

    Starkenburg , E., Martin , N., Youakim , K., et al. 2017, , 471, 2587, 10.1093/mnras/stx1068

    2017 doi

  68. [77]

    2006, , 132, 1645, 10.1086/506564

    Steinmetz , M., Zwitter , T., Siebert , A., et al. 2006, , 132, 1645, 10.1086/506564

    2006 doi

  69. [78]

    Stephenson , C. B. 1973, Publications of the Warner & Swasey Observatory

    1973

  70. [79]

    u eros , M., & Covarrubias , R. 2002, in Astronomical Society of the Pacific Conference Series, Vol. 261, The Physics of Cataclysmic Variables and Related Objects, ed. B. T. G \

    Szkody , P., Anderson , S., Ag \"u eros , M., & Covarrubias , R. 2002, in Astronomical Society of the Pacific Conference Series, Vol. 261, The Physics of Cataclysmic Variables and Related Objects, ed. B. T. G \"a nsicke , K. Beuermann , & K. Reinsch , 297

    2002

  71. [80]

    2017, , 228, 24, 10.3847/1538-4365/228/2/24

    Traven , G., Matijevi c , G., Zwitter , T., et al. 2017, , 228, 24, 10.3847/1538-4365/228/2/24

    2017 doi

  72. [81]

    2008, Journal of Machine Learning Research, 9, 2579

    van der Maaten, L., & Hinton, G. 2008, Journal of Machine Learning Research, 9, 2579. http://www.jmlr.org/papers/v9/vandermaaten08a.html

    2008

  73. [82]

    2019, , 483, 3196, 10.1093/mnras/sty3155

    C otar , K., Zwitter , T., Kos , J., et al. 2019, , 483, 3196, 10.1093/mnras/sty3155

    2019 doi

  74. [83]

    Wallerstein , G., & Knapp , G. R. 1998, , 36, 369, 10.1146/annurev.astro.36.1.369

    1998 doi

  75. [84]

    D., Placco , V

    Whitten , D. D., Placco , V. M., Beers , T. C., et al. 2021, , 912, 147, 10.3847/1538-4357/abee7e

    2021 doi

  76. [85]

    L., Eisenhardt , P

    Wright , E. L., Eisenhardt , P. R. M., Mainzer , A. K., et al. 2010, , 140, 1868, 10.1088/0004-6256/140/6/1868

    2010 doi

  77. [86]

    2015, , 452, 158, 10.1093/mnras/stv1210

    Yang , T., & Li , X. 2015, , 452, 158, 10.1093/mnras/stv1210

    2015 doi

  78. [87]

    P., Koposov , S

    Yao , Y., Ji , A. P., Koposov , S. E., & Limberg , G. 2024, , 527, 10937, 10.1093/mnras/stad3775

    2024 doi

  79. [88]

    G., Adelman , J., Anderson , John E., J., et al

    York , D. G., Adelman , J., Anderson , John E., J., et al. 2000, , 120, 1579, 10.1086/301513

    2000 doi

  80. [89]

    2009, , 508, 909, 10.1051/0004-6361/200912843

    Zamora , O., Abia , C., Plez , B., Dom \' nguez , I., & Cristallo , S. 2009, , 508, 909, 10.1051/0004-6361/200912843

    2009 doi

  81. [90]

    S., Li , Y., & Bi , S

    Zhang , X., Jeffery , C. S., Li , Y., & Bi , S. 2020, , 889, 33, 10.3847/1538-4357/ab5e89

    2020 doi

  82. [91]

    2012, Research in Astronomy and Astrophysics, 12, 723, 10.1088/1674-4527/12/7/002

    Zhao , G., Zhao , Y.-H., Chu , Y.-Q., Jing , Y.-P., & Deng , L.-C. 2012, Research in Astronomy and Astrophysics, 12, 723, 10.1088/1674-4527/12/7/002

    2012 doi

Pith tools

Reviewed August 16, 2026 · model on record in the stance chip above.