pith. sign in

arxiv: 2606.04567 · v1 · pith:XJX63HRWnew · submitted 2026-06-03 · 🌌 astro-ph.GA · astro-ph.SR

The GALAH Survey: Neutron-Capture Elemental Abundances for 350,000 Gaia-RVS Spectra and the Chemodynamics of Accreted Structures

Pradosh Barun Das , Daniel B. Zucker , Aldo Mura-Guzm\'an , Nicholas W. Borsato , Gayandhi M. De Silva , Sven Buder , Diane Feuillet , Thomas Nordlander
show 10 more authors
Melissa K. Ness Sarah L. Martell Janez Kos Joss Bland-Hawthorn Ken C. Freeman Andrew R. Casey Geraint F. Lewis Dennis Stello Richard de Grijs the GALAH Collaboration
This is my paper

Pith reviewed 2026-06-28 05:46 UTC · model grok-4.3

classification 🌌 astro-ph.GA astro-ph.SR
keywords GALAH SurveyGaia RVSneutron-capture elementsGaia-Sausage-Enceladuschemical taggingstellar abundancesMilky Way chemodynamicslogistic regression
0
0 comments X

The pith

A logistic regression classifier on chemical abundances from 357k Gaia RVS spectra identifies Gaia-Sausage-Enceladus members with persistent distinctive patterns.

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

The paper derives eleven stellar labels including neutron-capture abundances for 357,415 red giant stars from medium-resolution Gaia DR3 RVS spectra by training The Cannon on GALAH DR4 labels. It then trains a logistic regression classifier via MCMC on a small reference sample of GSE members and comparison stars to assign high membership probabilities using abundances alone. The resulting candidates show distinctive ratios in [Ca/Ti], [Ti/Ce], and [Nd/Zr] that remain consistent once independent kinematic constraints are applied. A sympathetic reader would care because the work shows how chemical information can trace the Milky Way's accretion history at the scale of hundreds of thousands of stars.

Core claim

Using The Cannon trained on 2747 common giants, we predict stellar parameters and abundances including [Zr/Fe], [Ce/Fe], [Nd/Fe] for 357,415 RVS stars. A logistic regression classifier optimised via MCMC and trained on a reference GSE sample identifies stars with high membership probabilities based on chemical abundances alone, with candidates exhibiting distinctive patterns in [Ca/Ti], [Ti/Ce], and [Nd/Zr] that hold after kinematic constraints are applied.

What carries the argument

Logistic regression classifier optimised via Markov Chain Monte Carlo sampling and trained on chemical abundances to identify accreted stars.

If this is right

  • The derived abundances enable large-scale chemodynamic studies of the Milky Way using neutron-capture elements.
  • Chemical signatures alone can identify accreted structures with patterns that survive kinematic filtering.
  • The data-driven framework extracts detailed abundances from medium-resolution spectra at the scale of hundreds of thousands of stars.
  • Neutron-capture ratios such as [Ti/Ce] and [Nd/Zr] provide additional leverage for distinguishing accreted populations.

Where Pith is reading between the lines

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

  • Similar classifiers could be trained on reference samples for other known accretion events such as the Helmi streams.
  • The method could be tested against simulated merger remnants to check how well abundance patterns separate from in-situ stars.
  • Extending the training to include more elements or photometric data might tighten membership probabilities further.

Load-bearing premise

The small reference sample of GSE members and comparison stars is sufficiently pure, representative, and chemically distinct from other Milky Way populations.

What would settle it

Higher-resolution follow-up spectroscopy of the high-probability candidates that measures the same abundance ratios and checks consistency with the training sample would confirm or refute the chemical classification.

Figures

Figures reproduced from arXiv: 2606.04567 by Aldo Mura-Guzm\'an, Andrew R. Casey, Daniel B. Zucker, Dennis Stello, Diane Feuillet, Gayandhi M. De Silva, Geraint F. Lewis, Janez Kos, Joss Bland-Hawthorn, Ken C. Freeman, Melissa K. Ness, Nicholas W. Borsato, Pradosh Barun Das, Richard de Grijs, Sarah L. Martell, Sven Buder, the GALAH collaboration, Thomas Nordlander.

Figure 1
Figure 1. Figure 1: Kiel Diagram showing the GALAH values for log 𝑔 versus 𝑇eff, colour coded with [Fe/H] for the 2747 stars selected in the training sample. The representative error bar in the lower-right corner indicates the median uncertainties (Δ𝑇eff ∼ 67 K and Δlog 𝑔 ∼ 0.1 dex) estimated in GALAH for the training sample. a comprehensive coverage that is essential for robust predictions in subsequent test samples (see [P… view at source ↗
Figure 2
Figure 2. Figure 2: One-to-one relations for The Cannon-predicted stellar labels for the 2747 Gaia-RVS spectra of giants in the training sample. The horizontal axis indicates The Cannon predictions of the stellar parameters and abundances, and the vertical axis shows the literature values from GALAH DR4. The black dashed line indicates the one-to-one relation, while the dotted lines mark ±2𝜎, where 𝜎 is the standard deviation… view at source ↗
Figure 3
Figure 3. Figure 3: Residual values for The Cannon-predicted stellar labels (see [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Comparison of two stars in our sample with similar stellar parameters and Δ[Nd/Fe] ∼ 0.60 dex. The Gaia-RVS spectra correspond to a star with a higher Nd abundance predicted by The Cannon, [Nd/Fe] = 0.61 dex (blue; Gaia ID: 6809018744787777024) and another with [Nd/Fe] = −0.02 dex (red; Gaia ID: 51670728611455360). The Nd abundances and stellar parameters and SNR (𝑇eff/log 𝑔/[Fe/H]/SNR) of both stars are s… view at source ↗
Figure 5
Figure 5. Figure 5: Cross-correlation detections of Ca II, Zr I, Ce II, and Nd II. The cross-correlation signal is shown as a black line with its 1𝜎 confidence interval marked as a grey region. plied using the values 𝑈⊙ = 11.1 km s−1 , 𝑉⊙ = 12.24 km s−1 , and 𝑊⊙ = 7.25 km s−1 (Bennett & Bovy 2019). These corrections were used to transform heliocentric velocities to the Galactocentric frame, after which orbits were integrated … view at source ↗
Figure 6
Figure 6. Figure 6: Posterior distributions and pairwise marginals for the MCMC model parameters (refer to Sec. 5.1). Each panel shows samples from the posterior distributions of the abundance ratio coefficients (𝛽 parameters) inferred from the model. The diagonal panels display the one-dimensional marginal posterior distributions for each parameter, with the mean (𝜇) and standard deviation (𝜎) indicated, together with the 95… view at source ↗
Figure 7
Figure 7. Figure 7: Distributions and pairwise correlations of the abundance ra￾tios—[Ca/Ti], [Ti/Ce], and [Nd/Zr]—for four distinct stellar populations: 2289 probable GSE candidate stars with 𝑃mem > 55% (red); halo stars (blue), thick disc stars (green), and a cross-match of 86 GSE stars from (Feuillet et al. 2021) with our catalogue of 314,010 Gaia-RVS stars (black) satisfying flag_cannon = 0. Diagonal panels display Kernel… view at source ↗
Figure 8
Figure 8. Figure 8: Kernel density estimates of the selected chemical abundance ratios—[Ca/Ti], [Ti/Ce], and [Nd/Zr]—for 2289 GSE candidate stars (red), and the Milky Way halo stars (blue) restricted to [Fe/H] < −0.8, [Ti/Fe] > 0.25 dex, 0.5 < 𝑒 < 0.8, and |𝐿𝑧 | < 1500 kpc km s−1 . For each abundance ratio, the peak of the distribution (dashed line) and 1𝜎 interval (16th–84th percentile, shaded region) are indicated. Annotati… view at source ↗
Figure 9
Figure 9. Figure 9: Each panel shows the normalized histogram of a given element ratio ([Fe/H], [Ti/Fe], [Ca/Fe], [Ni/Fe], [Si/Fe], [Nd/Fe], [Zr/Fe], [Ce/Fe]) for four samples of GSE stars: 2289 GSE candidates selected chemically in our work (𝑃mem > 55%) marked with blue; a subset of 286 GSE candidates selected chemically in our sample (𝑃mem > 55%) and satisfying the dynamical conditions from Feuillet et al. (2021) [−500 ≤ 𝐿𝑧… view at source ↗
Figure 10
Figure 10. Figure 10: Kinematic distribution of red giant stars analyzed with The Cannon. Left: Total orbital energy (𝐸) versus angular momentum (𝐿𝑧 ). The golden coloured-points show the RVS sample of 314,010 RVS giants (flag_cannon = 0). Star-shaped markers denote probable GSE members with 𝑃mem > 55%, selected using the abundance-based MCMC model. Diamond-shaped markers show the subset of stars that also satisfy the dynamica… view at source ↗
read the original abstract

We present a comprehensive data-driven spectroscopic analysis of 357,415 red giant stars using Gaia DR3 Radial Velocity Spectrometer (RVS) spectra (8460-8700 A; $R\approx11,500$), aimed at deriving precise stellar parameters and elemental abundances (collectively referred to as stellar labels). We employ The Cannon, a generative model based on 2747 giants in common with GALAH DR4, adopting GALAH labels ($R\approx28,000$) for training. The resulting model predicts eleven stellar labels for RVS giants: effective temperature ($T_{\rm eff}$), surface gravity ($\log g$), projected rotational velocity ($v\sin i$), and abundances of [Fe/H], [Ca/Fe], [Si/Fe], [Ni/Fe], [Ti/Fe], as well as the neutron-capture elements [Zr/Fe], [Ce/Fe], and [Nd/Fe]. Building on these results, we develop a probabilistic framework to chemically identify debris from the Gaia-Sausage-Enceladus (GSE) accretion event. A logistic regression classifier, optimised via Markov Chain Monte Carlo sampling and trained on a small reference sample of GSE members and comparison stars, identifies stars with high GSE membership probabilities based solely on their chemical abundances, with the resulting candidates exhibiting distinctive abundance-ratio patterns, including [Ca/Ti], [Ti/Ce], and [Nd/Zr]. Applying independent kinematic constraints yields a robust sample of GSE candidates, demonstrating that the characteristic chemical signatures remain consistent after applying these constraints. This work demonstrates the power of data-driven analysis techniques to extract detailed chemical information from medium-resolution spectra and establishes a framework for tracing Galactic accretion events using chemical abundances.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 2 minor

Summary. The paper applies The Cannon, trained on 2747 GALAH DR4 giants, to derive 11 stellar labels (including [Zr/Fe], [Ce/Fe], [Nd/Fe]) from 357,415 Gaia RVS spectra. It then trains a logistic regression classifier via MCMC on chemical abundances from a small reference set of GSE members and comparison stars to assign GSE membership probabilities, reporting that high-probability candidates show distinctive ratios ([Ca/Ti], [Ti/Ce], [Nd/Zr]) that persist after independent kinematic filtering.

Significance. If the classifier generalizes, the work would deliver one of the largest chemically tagged GSE samples to date and demonstrate that medium-resolution RVS spectra can yield usable neutron-capture abundances for chemodynamic studies of accretion events. The data-driven pipeline itself is a useful technical contribution for large spectroscopic surveys.

major comments (3)
  1. [GSE classifier description (near end of abstract and corresponding methods/results)] The section describing the logistic regression classifier provides no quantitative details on the size of the GSE reference sample, its selection criteria (kinematic, abundance, or otherwise), purity estimates, or chemical distinctness metrics relative to other Milky Way populations. This information is required to assess whether the learned decision boundary generalizes or reproduces the input selection.
  2. [Results on abundance-ratio patterns and kinematic filtering] No cross-validation performance, confusion matrix, or comparison against an independent GSE catalog is reported for the classifier. Without these, the claim that the [Ca/Ti], [Ti/Ce], and [Nd/Zr] patterns are robust GSE signatures (rather than artifacts of the training set) cannot be evaluated.
  3. [GSE identification framework] The abstract states that the classifier is trained 'based solely on their chemical abundances,' yet the reference sample selection method is not shown; if that sample was itself defined with kinematic cuts, the subsequent 'independent kinematic constraints' test is not fully independent and the circularity risk noted in the stress-test applies directly.
minor comments (2)
  1. [Data and model description] The training-set size (2747 stars) and label list are stated, but the wavelength range and resolution of the RVS spectra are given only in the abstract; a dedicated table or paragraph in the methods would improve clarity.
  2. [Abundance results] Notation for the neutron-capture ratios (e.g., [Nd/Zr]) should be defined explicitly the first time they appear, including whether they are [X/Fe] or [X/Y] quantities.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for their constructive and detailed comments, which highlight important areas for clarification and validation in our GSE identification framework. We address each major comment below and will revise the manuscript to incorporate the requested information and analyses.

read point-by-point responses

  1. Referee: [GSE classifier description (near end of abstract and corresponding methods/results)] The section describing the logistic regression classifier provides no quantitative details on the size of the GSE reference sample, its selection criteria (kinematic, abundance, or otherwise), purity estimates, or chemical distinctness metrics relative to other Milky Way populations. This information is required to assess whether the learned decision boundary generalizes or reproduces the input selection.

    Authors: We agree that quantitative details on the reference sample are necessary. The current manuscript describes it only as 'small' without specifics. In the revised version, we will expand the methods section to report the exact sample size, full selection criteria (including any kinematic or abundance cuts from the literature), purity estimates, and metrics of chemical distinctness relative to other populations. This will enable readers to evaluate generalization. revision: yes

  2. Referee: [Results on abundance-ratio patterns and kinematic filtering] No cross-validation performance, confusion matrix, or comparison against an independent GSE catalog is reported for the classifier. Without these, the claim that the [Ca/Ti], [Ti/Ce], and [Nd/Zr] patterns are robust GSE signatures (rather than artifacts of the training set) cannot be evaluated.

    Authors: We acknowledge the absence of these validation metrics. We will add cross-validation results for the logistic regression, include a confusion matrix, and compare our high-probability candidates against an independent GSE catalog (where overlaps exist) in the revised manuscript. These additions will support the robustness of the reported abundance patterns. revision: yes

  3. Referee: [GSE identification framework] The abstract states that the classifier is trained 'based solely on their chemical abundances,' yet the reference sample selection method is not shown; if that sample was itself defined with kinematic cuts, the subsequent 'independent kinematic constraints' test is not fully independent and the circularity risk noted in the stress-test applies directly.

    Authors: The reference sample draws from literature GSE members that were originally identified with both kinematic and chemical information, though the classifier uses only abundances. We will revise the text to explicitly detail the reference sample construction and discuss the independence of the subsequent kinematic test, including limitations and additional stress-tests to mitigate circularity concerns. revision: partial

Circularity Check

0 steps flagged

No significant circularity

full rationale

The abstract describes training a logistic regression classifier on a reference sample of GSE members and comparison stars, then applying it to predict membership probabilities in the 357k-star catalog based on chemical abundances, followed by independent kinematic validation. No equations, self-citations, or descriptions are provided that reduce the output probabilities or abundance patterns to the training inputs by construction. The workflow is a standard supervised learning pipeline with an external check, making the derivation self-contained against the given text.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract-only review supplies no explicit free parameters, axioms, or invented entities; the logistic regression implicitly treats the reference sample abundances as ground truth without stated uncertainty propagation.

pith-pipeline@v0.9.1-grok · 5951 in / 1143 out tokens · 34499 ms · 2026-06-28T05:46:32.720791+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

83 extracted references · 69 canonical work pages · 9 internal anchors

  1. [12]

    2018 , howpublished =

    James Bradbury and Roy Frostig and Peter Hawkins and Matthew James Johnson and Chris Leary and Dougal Maclaurin and Skye Wanderman-Milne , title =. 2018 , howpublished =

    2018

  2. [25]

    VizieR Online Data Catalog: Gaia RVS stellar labels using The Cannon (Das+, 2025)

    2025

  3. [36]

    The Messenger , year = 2012, month = mar, volume =

    The Gaia-ESO Public Spectroscopic Survey. The Messenger , year = 2012, month = mar, volume =

    2012

  4. [55]

    SciPy , year=

    Data Structures for Statistical Computing in Python , author=. SciPy , year=

  5. [64]

    The Messenger , year = 2013, month = dec, volume =

    The Gaia-ESO Large Public Spectroscopic Survey. The Messenger , year = 2013, month = dec, volume =

    2013

  6. [74]

    Astronomical Data Analysis Software and Systems XIV , year = 2005, editor =

    TOPCAT & STIL: Starlink Table/VOTable Processing Software. Astronomical Data Analysis Software and Systems XIV , year = 2005, editor =

    2005

  7. [75]

    TOPCAT: Tool for OPerations on Catalogues And Tables

  8. [77]

    Astropy Collaboration et al., 2013, @doi [ ] 10.1051/0004-6361/201322068 , https://ui.adsabs.harvard.edu/abs/2013A&A...558A..33A 558, A33

    work page doi:10.1051/0004-6361/201322068 2013

  9. [78]

    Astropy Collaboration et al., 2018, @doi [ ] 10.3847/1538-3881/aabc4f , https://ui.adsabs.harvard.edu/abs/2018AJ....156..123A 156, 123

    work page doi:10.3847/1538-3881/aabc4f 2018

  10. [79]

    Astropy Collaboration et al., 2022, @doi [ ] 10.3847/1538-4357/ac7c74 , https://ui.adsabs.harvard.edu/abs/2022ApJ...935..167A 935, 167

    work page internal anchor Pith review doi:10.3847/1538-4357/ac7c74 2022

  11. [80]

    W., Koposov S

    Belokurov V., Erkal D., Evans N. W., Koposov S. E., Deason A. J., 2018, @doi [ ] 10.1093/mnras/sty982 , https://ui.adsabs.harvard.edu/abs/2018MNRAS.478..611B 478, 611

    work page doi:10.1093/mnras/sty982 2018

  12. [81]

    Bennett M., Bovy J., 2019, @doi [ ] 10.1093/mnras/sty2813 , https://ui.adsabs.harvard.edu/abs/2019MNRAS.482.1417B 482, 1417

    work page internal anchor Pith review doi:10.1093/mnras/sty2813 2019

  13. [82]

    Berni L., Palla M., Magrini L., Spina L., 2026, @doi [ ] 10.1051/0004-6361/202558043 , https://ui.adsabs.harvard.edu/abs/2026A&A...707A.242B 707, A242

    work page doi:10.1051/0004-6361/202558043 2026

  14. [83]

    The Convergence of Markov chain Monte Carlo Methods: From the Metropolis method to Hamiltonian Monte Carlo

    Betancourt M., 2017, @doi [arXiv e-prints] 10.48550/arXiv.1706.01520 , https://ui.adsabs.harvard.edu/abs/2017arXiv170601520B p. arXiv:1706.01520

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.1706.01520 2017

  15. [84]

    Pyro: Deep Universal Probabilistic Programming

    Bingham E., et al., 2018, @doi [arXiv e-prints] 10.48550/arXiv.1810.09538 , https://ui.adsabs.harvard.edu/abs/2018arXiv181009538B p. arXiv:1810.09538

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.1810.09538 2018

  16. [85]

    W., Hoeijmakers H

    Borsato N. W., Hoeijmakers H. J., Prinoth B., Thorsbro B., Forsberg R., Kitzmann D., Jones K., Heng K., 2023, @doi [ ] 10.1051/0004-6361/202245121 , https://ui.adsabs.harvard.edu/abs/2023A&A...673A.158B 673, A158

    work page doi:10.1051/0004-6361/202245121 2023

  17. [86]

    Bovy J., 2015, @doi [ ] 10.1088/0067-0049/216/2/29 , https://ui.adsabs.harvard.edu/abs/2015ApJS..216...29B 216, 29

    work page internal anchor Pith review doi:10.1088/0067-0049/216/2/29 2015

  18. [87]

    Bovy J., Rix H.-W., 2013, @doi [ ] 10.1088/0004-637X/779/2/115 , https://ui.adsabs.harvard.edu/abs/2013ApJ...779..115B 779, 115

    work page doi:10.1088/0004-637x/779/2/115 2013

  19. [88]

    J., Leary C., Maclaurin D., Wanderman-Milne S., 2018, JAX: Composable transformations of Python+NumPy programs, https://github.com/google/jax

    Bradbury J., Frostig R., Hawkins P., Johnson M. J., Leary C., Maclaurin D., Wanderman-Milne S., 2018, JAX: Composable transformations of Python+NumPy programs, https://github.com/google/jax

    2018

  20. [89]

    E., ed., Society of Photo-Optical Instrumentation Engineers (SPIE) Conference Series Vol

    Brzeski J., Case S., Gers L., 2011, in Hatheway A. E., ed., Society of Photo-Optical Instrumentation Engineers (SPIE) Conference Series Vol. 8125, Optomechanics 2011: Innovations and Solutions. p. 812504, @doi 10.1117/12.896389

    work page doi:10.1117/12.896389 2011

  21. [90]

    Buder S., et al., 2018, @doi [ ] 10.1093/mnras/sty1281 , https://ui.adsabs.harvard.edu/abs/2018MNRAS.478.4513B 478, 4513

    work page doi:10.1093/mnras/sty1281 2018

  22. [91]

    Buder S., et al., 2021, @doi [ ] 10.1093/mnras/stab1242 , https://ui.adsabs.harvard.edu/abs/2021MNRAS.506..150B 506, 150

    work page doi:10.1093/mnras/stab1242 2021

  23. [92]

    Buder S., et al., 2022, @doi [ ] 10.1093/mnras/stab3504 , https://ui.adsabs.harvard.edu/abs/2022MNRAS.510.2407B 510, 2407

    work page doi:10.1093/mnras/stab3504 2022

  24. [93]

    Buder S., et al., 2025, @doi [ ] 10.1017/pasa.2025.26 , https://ui.adsabs.harvard.edu/abs/2025PASA...42...51B 42, e051

    work page doi:10.1017/pasa.2025.26 2025

  25. [94]

    Carrillo A., Hawkins K., Jofr \'e P., de Brito Silva D., Das P., Lucey M., 2022, @doi [ ] 10.1093/mnras/stac518 , https://ui.adsabs.harvard.edu/abs/2022MNRAS.513.1557C 513, 1557

    work page doi:10.1093/mnras/stac518 2022

  26. [95]

    The Cannon 2: A data-driven model of stellar spectra for detailed chemical abundance analyses

    Casey A. R., Hogg D. W., Ness M., Rix H.-W., Ho A. Q. Y., Gilmore G., 2016, @doi [arXiv e-prints] 10.48550/arXiv.1603.03040 , https://ui.adsabs.harvard.edu/abs/2016arXiv160303040C p. arXiv:1603.03040

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.1603.03040 2016

  27. [96]

    New Grids of ATLAS9 Model Atmospheres

    Castelli F., Kurucz R. L., 2003, in Piskunov N., Weiss W. W., Gray D. F., eds, IAU Symposium Vol. 210, Modelling of Stellar Atmospheres. p. A20 ( @eprint arXiv astro-ph/0405087 ), @doi 10.48550/arXiv.astro-ph/0405087

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.astro-ph/0405087 2003

  28. [97]

    R., 1958, @doi [Journal of the Royal Statistical Society: Series B (Methodological)] 10.1111/j.2517-6161.1958.tb00292.x , 20, 215

    Cox D. R., 1958, @doi [Journal of the Royal Statistical Society: Series B (Methodological)] 10.1111/j.2517-6161.1958.tb00292.x , 20, 215

    work page doi:10.1111/j.2517-6161.1958.tb00292.x 1958

  29. [98]

    Cropper M., et al., 2018, @doi [ ] 10.1051/0004-6361/201832763 , https://ui.adsabs.harvard.edu/abs/2018A&A...616A...5C 616, A5

    work page doi:10.1051/0004-6361/201832763 2018

  30. [99]

    Das P., Hawkins K., Jofr \'e P., 2020, @doi [ ] 10.1093/mnras/stz3537 , https://ui.adsabs.harvard.edu/abs/2020MNRAS.493.5195D 493, 5195

    work page doi:10.1093/mnras/stz3537 2020

  31. [100]

    B., et al., 2025a, VizieR Online Data Catalog: Gaia RVS stellar labels using The Cannon (Das+, 2025) , VizieR On-line Data Catalog: J/MNRAS/538/605

    Das P. B., et al., 2025a, VizieR Online Data Catalog: Gaia RVS stellar labels using The Cannon (Das+, 2025) , VizieR On-line Data Catalog: J/MNRAS/538/605. Originally published in: 2025MNRAS.538..605D

    2025

  32. [101]

    B., et al., 2025b, @doi [ ] 10.1093/mnras/staf169 , https://ui.adsabs.harvard.edu/abs/2025MNRAS.538..605D 538, 605

    Das P. B., et al., 2025b, @doi [ ] 10.1093/mnras/staf169 , https://ui.adsabs.harvard.edu/abs/2025MNRAS.538..605D 538, 605

    work page doi:10.1093/mnras/staf169

  33. [102]

    M., et al., 2015, @doi [ ] 10.1093/mnras/stv327 , https://ui.adsabs.harvard.edu/abs/2015MNRAS.449.2604D 449, 2604

    De Silva G. M., et al., 2015, @doi [ ] 10.1093/mnras/stv327 , https://ui.adsabs.harvard.edu/abs/2015MNRAS.449.2604D 449, 2604

    work page doi:10.1093/mnras/stv327 2015

  34. [103]

    Ernandes H., Feuillet D., Feltzing S., Sk \'u lad \'o ttir \'A ., 2024, @doi [ ] 10.1051/0004-6361/202450827 , https://ui.adsabs.harvard.edu/abs/2024A&A...691A.333E 691, A333

    work page doi:10.1051/0004-6361/202450827 2024

  35. [104]

    Ernandes H., Sk \'u lad \'o ttir \'A ., Feltzing S., Feuillet D., 2025, @doi [ ] 10.1051/0004-6361/202557149 , https://ui.adsabs.harvard.edu/abs/2025A&A...703A.256E 703, A256

    work page doi:10.1051/0004-6361/202557149 2025

  36. [105]

    K., Feltzing S., Sahlholdt C

    Feuillet D. K., Feltzing S., Sahlholdt C. L., Casagrande L., 2020, @doi [ ] 10.1093/mnras/staa1888 , https://ui.adsabs.harvard.edu/abs/2020MNRAS.497..109F 497, 109

    work page doi:10.1093/mnras/staa1888 2020

  37. [106]

    K., Sahlholdt C

    Feuillet D. K., Sahlholdt C. L., Feltzing S., Casagrande L., 2021, @doi [ ] 10.1093/mnras/stab2614 , https://ui.adsabs.harvard.edu/abs/2021MNRAS.508.1489F 508, 1489

    work page doi:10.1093/mnras/stab2614 2021

  38. [107]

    K., Yong D., Karakas A

    Fishlock C. K., Yong D., Karakas A. I., Alves-Brito A., Mel \'e ndez J., Nissen P. E., Kobayashi C., Casey A. R., 2017, @doi [ ] 10.1093/mnras/stx047 , https://ui.adsabs.harvard.edu/abs/2017MNRAS.466.4672F 466, 4672

    work page doi:10.1093/mnras/stx047 2017

  39. [108]

    Gaia Collaboration et al., 2016, @doi [ ] 10.1051/0004-6361/201629272 , https://ui.adsabs.harvard.edu/abs/2016A&A...595A...1G 595, A1

    work page doi:10.1051/0004-6361/201629272 2016

  40. [109]

    Gaia Collaboration et al., 2018, @doi [ ] 10.1051/0004-6361/201833051 , https://ui.adsabs.harvard.edu/abs/2018A&A...616A...1G 616, A1

    work page internal anchor Pith review doi:10.1051/0004-6361/201833051 2018

  41. [110]

    Gaia Collaboration et al., 2023a, @doi [ ] 10.1051/0004-6361/202243940 , https://ui.adsabs.harvard.edu/abs/2023A&A...674A...1G 674, A1

    work page doi:10.1051/0004-6361/202243940

  42. [111]

    Gaia Collaboration et al., 2023b, @doi [ ] 10.1051/0004-6361/202243511 , https://ui.adsabs.harvard.edu/abs/2023A&A...674A..38G 674, A38

    work page doi:10.1051/0004-6361/202243511

  43. [112]

    Gilmore G., et al., 2012, The Messenger, https://ui.adsabs.harvard.edu/abs/2012Msngr.147...25G 147, 25

    2012

  44. [113]

    Guiglion G., et al., 2024, @doi [ ] 10.1051/0004-6361/202347122 , https://ui.adsabs.harvard.edu/abs/2024A&A...682A...9G 682, A9

    work page doi:10.1051/0004-6361/202347122 2024

  45. [114]

    R., Millman, K

    Harris C. R., et al., 2020, @doi [ ] 10.1038/s41586-020-2649-2 , https://ui.adsabs.harvard.edu/abs/2020Natur.585..357H 585, 357

    work page doi:10.1038/s41586-020-2649-2 2020

  46. [115]

    Hawkins K., Jofr \'e P., Masseron T., Gilmore G., 2015, @doi [ ] 10.1093/mnras/stv1586 , https://ui.adsabs.harvard.edu/abs/2015MNRAS.453..758H 453, 758

    work page doi:10.1093/mnras/stv1586 2015

  47. [116]

    S., Ramsay S

    Heijmans J., et al., 2012, in McLean I. S., Ramsay S. K., Takami H., eds, Society of Photo-Optical Instrumentation Engineers (SPIE) Conference Series Vol. 8446, Ground-based and Airborne Instrumentation for Astronomy IV. p. 84460W, @doi 10.1117/12.925806

    work page doi:10.1117/12.925806 2012

  48. [117]

    H., Massari D., Veljanoski J., Brown A

    Helmi A., Babusiaux C., Koppelman H. H., Massari D., Veljanoski J., Brown A. G. A., 2018, @doi [ ] 10.1038/s41586-018-0625-x , https://ui.adsabs.harvard.edu/abs/2018Natur.563...85H 563, 85

    work page doi:10.1038/s41586-018-0625-x 2018

  49. [118]

    Ho A. Y. Q., et al., 2017a, @doi [ ] 10.3847/1538-4357/836/1/5 , https://ui.adsabs.harvard.edu/abs/2017ApJ...836....5H 836, 5

    work page doi:10.3847/1538-4357/836/1/5

  50. [119]

    Ho A. Y. Q., Rix H.-W., Ness M. K., Hogg D. W., Liu C., Ting Y.-S., 2017b, @doi [ ] 10.3847/1538-4357/aa6db3 , https://ui.adsabs.harvard.edu/abs/2017ApJ...841...40H 841, 40

    work page doi:10.3847/1538-4357/aa6db3

  51. [120]

    K., 2025, @doi [ ] 10.1093/mnrasl/slaf093 , https://ui.adsabs.harvard.edu/abs/2025MNRAS.544L..64H 544, L64

    Horta D., Ness M. K., 2025, @doi [ ] 10.1093/mnrasl/slaf093 , https://ui.adsabs.harvard.edu/abs/2025MNRAS.544L..64H 544, L64

    work page doi:10.1093/mnrasl/slaf093 2025

  52. [121]

    2013, Applied Logistic Regression: Third Edition (wiley), doi: 10.1002/9781118548387

    Hosmer D., Lemeshow S., Sturdivant R., 2013, Applied Logistic Regression: Third Edition. wiley, @doi 10.1002/9781118548387

    work page doi:10.1002/9781118548387 2013

  53. [122]

    Computing in Science and Engineering , keywords =

    Hunter J. D., 2007, @doi [Computing in Science and Engineering] 10.1109/MCSE.2007.55 , https://ui.adsabs.harvard.edu/abs/2007CSE.....9...90H 9, 90

    work page doi:10.1109/mcse.2007.55 2007

  54. [123]

    Koppelman H., Helmi A., Veljanoski J., 2018, @doi [ ] 10.3847/2041-8213/aac882 , https://ui.adsabs.harvard.edu/abs/2018ApJ...860L..11K 860, L11

    work page doi:10.3847/2041-8213/aac882 2018

  55. [124]

    H., Helmi A., Massari D., Price-Whelan A

    Koppelman H. H., Helmi A., Massari D., Price-Whelan A. M., Starkenburg T. K., 2019, @doi [ ] 10.1051/0004-6361/201936738 , https://ui.adsabs.harvard.edu/abs/2019A&A...631L...9K 631, L9

    work page doi:10.1051/0004-6361/201936738 2019

  56. [125]

    R., Tout, C

    Lewis I. J., et al., 2002, @doi [ ] 10.1046/j.1365-8711.2002.05333.x , https://ui.adsabs.harvard.edu/abs/2002MNRAS.333..279L 333, 279

    work page doi:10.1046/j.1365-8711.2002.05333.x 2002

  57. [126]

    Luo A.-L., et al., 2015, @doi [Research in Astronomy and Astrophysics] 10.1088/1674-4527/15/8/002 , https://ui.adsabs.harvard.edu/abs/2015RAA....15.1095L 15, 1095

    work page doi:10.1088/1674-4527/15/8/002 2015

  58. [127]

    R., et al., 2017, @doi [ ] 10.3847/1538-3881/aa784d , https://ui.adsabs.harvard.edu/abs/2017AJ....154...94M 154, 94

    Majewski S. R., et al., 2017, @doi [ ] 10.3847/1538-3881/aa784d , https://ui.adsabs.harvard.edu/abs/2017AJ....154...94M 154, 94

    work page doi:10.3847/1538-3881/aa784d 2017

  59. [128]

    K., Buder S., Martell S

    Manea C., Hawkins K., Ness M. K., Buder S., Martell S. L., Zucker D. B., 2024, @doi [ ] 10.3847/1538-4357/ad58d9 , https://ui.adsabs.harvard.edu/abs/2024ApJ...972...69M 972, 69

    work page doi:10.3847/1538-4357/ad58d9 2024

  60. [129]

    L., et al., 2017, @doi [ ] 10.1093/mnras/stw2835 , https://ui.adsabs.harvard.edu/abs/2017MNRAS.465.3203M 465, 3203

    Martell S. L., et al., 2017, @doi [ ] 10.1093/mnras/stw2835 , https://ui.adsabs.harvard.edu/abs/2017MNRAS.465.3203M 465, 3203

    work page doi:10.1093/mnras/stw2835 2017

  61. [130]

    McKenzie M., et al., 2022, @doi [ ] 10.1093/mnras/stac2254 , https://ui.adsabs.harvard.edu/abs/2022MNRAS.516.3515M 516, 3515

    work page doi:10.1093/mnras/stac2254 2022

  62. [131]

    https://api.semanticscholar.org/CorpusID:13156234

    McKinney W., 2010, in SciPy. https://api.semanticscholar.org/CorpusID:13156234

    2010

  63. [132]

    F., Karakas A

    Monty S., Yong D., Marino A. F., Karakas A. I., McKenzie M., Grundahl F., Mura-Guzm \'a n A., 2023, @doi [ ] 10.1093/mnras/stac3040 , https://ui.adsabs.harvard.edu/abs/2023MNRAS.518..965M 518, 965

    work page doi:10.1093/mnras/stac3040 2023

  64. [133]

    P., Conroy C., Bonaca A., Johnson B

    Naidu R. P., Conroy C., Bonaca A., Johnson B. D., Ting Y.-S., Caldwell N., Zaritsky D., Cargile P. A., 2020, @doi [ ] 10.3847/1538-4357/abaef4 , https://ui.adsabs.harvard.edu/abs/2020ApJ...901...48N 901, 48

    work page doi:10.3847/1538-4357/abaef4 2020

  65. [134]

    P., et al., 2021, @doi [ ] 10.3847/1538-4357/ac2d2d , https://ui.adsabs.harvard.edu/abs/2021ApJ...923...92N 923, 92

    Naidu R. P., et al., 2021, @doi [ ] 10.3847/1538-4357/ac2d2d , https://ui.adsabs.harvard.edu/abs/2021ApJ...923...92N 923, 92

    work page doi:10.3847/1538-4357/ac2d2d 2021

  66. [135]

    Nandakumar G., et al., 2022, @doi [ ] 10.1093/mnras/stac873 , https://ui.adsabs.harvard.edu/abs/2022MNRAS.513..232N 513, 232

    work page doi:10.1093/mnras/stac873 2022

  67. [136]

    W., Rix H.-W., Ho A

    Ness M., Hogg D. W., Rix H.-W., Ho A. Y. Q., Zasowski G., 2015, @doi [ ] 10.1088/0004-637X/808/1/16 , https://ui.adsabs.harvard.edu/abs/2015ApJ...808...16N 808, 16

    work page doi:10.1088/0004-637x/808/1/16 2015

  68. [137]

    E., Schuster W

    Nissen P. E., Schuster W. J., 2010, @doi [ ] 10.1051/0004-6361/200913877 , https://ui.adsabs.harvard.edu/abs/2010A&A...511L..10N 511, L10

    work page doi:10.1051/0004-6361/200913877 2010

  69. [138]

    Composable Effects for Flexible and Accelerated Probabilistic Programming in NumPyro

    Phan D., Pradhan N., Jankowiak M., 2019, @doi [arXiv e-prints] 10.48550/arXiv.1912.11554 , https://ui.adsabs.harvard.edu/abs/2019arXiv191211554P p. arXiv:1912.11554

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.1912.11554 2019

  70. [139]

    M., Sneden C., Roederer I

    Placco V. M., Sneden C., Roederer I. U., Lawler J. E., Den Hartog E. A., Hejazi N., Maas Z., Bernath P., 2021, @doi [Research Notes of the American Astronomical Society] 10.3847/2515-5172/abf651 , https://ui.adsabs.harvard.edu/abs/2021RNAAS...5...92P 5, 92

    work page doi:10.3847/2515-5172/abf651 2021

  71. [140]

    Randich S., Gilmore G., Gaia-ESO Consortium 2013, The Messenger, https://ui.adsabs.harvard.edu/abs/2013Msngr.154...47R 154, 47

    2013

  72. [141]

    Recio-Blanco A., et al., 2016, @doi [ ] 10.1051/0004-6361/201425030 , https://ui.adsabs.harvard.edu/abs/2016A&A...585A..93R 585, A93

    work page doi:10.1051/0004-6361/201425030 2016

  73. [142]

    Recio-Blanco A., et al., 2023, @doi [ ] 10.1051/0004-6361/202243750 , https://ui.adsabs.harvard.edu/abs/2023A&A...674A..29R 674, A29

    work page doi:10.1051/0004-6361/202243750 2023

  74. [143]

    Sharma S., et al., 2018, @doi [ ] 10.1093/mnras/stx2582 , https://ui.adsabs.harvard.edu/abs/2018MNRAS.473.2004S 473, 2004

    work page doi:10.1093/mnras/stx2582 2018

  75. [144]

    Sharma S., et al., 2019, @doi [ ] 10.1093/mnras/stz2861 , https://ui.adsabs.harvard.edu/abs/2019MNRAS.490.5335S 490, 5335

    work page doi:10.1093/mnras/stz2861 2019

  76. [145]

    Sheinis A., et al., 2015, @doi [Journal of Astronomical Telescopes, Instruments, and Systems] 10.1117/1.JATIS.1.3.035002 , https://ui.adsabs.harvard.edu/abs/2015JATIS...1c5002S 1, 035002

    work page doi:10.1117/1.jatis.1.3.035002 2015

  77. [146]

    K., Mori A., Feltzing S., Lucchesi R

    Sk \'u lad \'o ttir \'A ., Ernandes H., Feuillet D. K., Mori A., Feltzing S., Lucchesi R. E. R., Di Matteo P., 2025, @doi [ ] 10.3847/2041-8213/addc66 , https://ui.adsabs.harvard.edu/abs/2025ApJ...986L..21S 986, L21

    work page doi:10.3847/2041-8213/addc66 2025

  78. [147]

    Sneden C., 1973, @doi [ ] 10.1086/152374 , https://ui.adsabs.harvard.edu/abs/1973ApJ...184..839S 184, 839

    work page doi:10.1086/152374 1973

  79. [148]

    J., Gallino R., 2008, @doi [ ] 10.1146/annurev.astro.46.060407.145207 , https://ui.adsabs.harvard.edu/abs/2008ARA&A..46..241S 46, 241

    Sneden C., Cowan J. J., Gallino R., 2008, @doi [ ] 10.1146/annurev.astro.46.060407.145207 , https://ui.adsabs.harvard.edu/abs/2008ARA&A..46..241S 46, 241

    work page doi:10.1146/annurev.astro.46.060407.145207 2008

  80. [149]

    Snellen I. A. G., de Kok R. J., de Mooij E. J. W., Albrecht S., 2010, @doi [ ] 10.1038/nature09111 , https://ui.adsabs.harvard.edu/abs/2010Natur.465.1049S 465, 1049

    work page doi:10.1038/nature09111 2010

Showing first 80 references.