Pith. sign in

REVIEW 3 major objections 5 minor 106 references

Updated morphological cuts identify galaxy mergers with purity matching machine-learning methods and better hold up at high redshift.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

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

T0 review · grok-4.5

2026-07-13 04:39 UTC pith:6V4NJ52S

load-bearing objection Clean re-optimisation of classic G-M20/G-C cuts that matches ML precision on the same mocks and is more robust than CNNs at high z, with the honest caveat that it only catches pre-mergers. the 3 major comments →

arxiv 2607.09209 v1 pith:6V4NJ52S submitted 2026-07-10 astro-ph.GA

Performance of morphological classifiers for galaxy mergers compared to current machine learning methods

Aidan P. Cotter , William J. pearson , Subhrata Dey , Berta Margalef-Bentabol , Alejandro Guzm\'an-Ortega , Vicente Rodriguez-Gomez This is my paper
classification astro-ph.GA
keywords galaxy mergersmorphological classificationGini coefficientM20 statisticconcentrationmachine learning comparisonIllustrisTNGHSC-SSP
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

This paper shows that simple non-parametric morphology statistics can select galaxy mergers about as cleanly as modern machine-learning classifiers, while remaining easier to apply and more stable when the data change. Using mock HSC-SSP images from IllustrisTNG, the authors optimise linear boundaries in the Gini–M20 and Gini–concentration planes and recover precisions of roughly 70 percent on held-out images, inside the range of six deep-learning and random-forest models trained on the same set. The same cuts keep their purity better than convolutional networks when transferred to a different simulation or to real HSC-SSP images, especially above redshift 0.5. The catch is that they mainly catch pre-coalescence systems; post-mergers look morphologically ordinary and therefore reduce completeness. The practical payoff is a set of transparent, low-cost classifiers that can be run on large surveys without the training and transfer costs of deep networks, provided new statistics are later found for the post-merger phase.

Core claim

Linear cuts optimised by MCMC in Gini–M20 and Gini–concentration space—G > (−0.267 ± 0.081)M20 + (0.143 ± 0.012) and G > (0.162 ± 0.048)C − (0.149 ± 0.12)—classify mergers with precisions of 69.5 percent and 72.3 percent on unseen IllustrisTNG mock HSC-SSP images, matching the purity range of six state-of-the-art ML models trained on identical data, while remaining more robust than CNN-based methods for 0.52 < z < 1.

What carries the argument

MCMC-optimised linear separators in two-dimensional Gini–M20 and Gini–concentration morphological space, using reciprocal accuracy as the likelihood and a validation set to avoid over-fitting.

Load-bearing premise

The simulation merger trees that supply the training labels (major mergers only, fixed time window around coalescence) are assumed to be an unbiased ground truth equally fair to both morphological cuts and the machine-learning models being compared.

What would settle it

Apply the published G–M20 and G–C cuts to an independent set of mock images whose merger trees are built with a different algorithm or a different mass-ratio and time window, and check whether the reported precisions and high-z robustness relative to the same ML models still hold.

Watch this falsifier — get emailed when new claim-graph text bears on it.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 5 minor

Summary. The paper optimises linear merger cuts in G–M20 and G–C space with MCMC (validation-set regularisation and truncated-Gaussian priors) on the same IllustrisTNG mock HSC-SSP training set used by Margalef-Bentabol et al. (2024). The resulting criteria G > (−0.267 ± 0.081)M20 + (0.143 ± 0.012) and G > (0.162 ± 0.048)C − (0.149 ± 0.12) yield precisions of 69.5 % and 72.3 % on held-out IllustrisTNG images—inside the 69.1–80.1 % range of six ML models trained on identical data—while recall remains low (36.4 % and 23.6 %). Subclass analysis shows the cuts select only pre-mergers; post- and ongoing mergers occupy the same locus as non-mergers. Transfer tests to Horizon-AGN and HSC-SSP, plus a robustness metric ρ, indicate that the morphological classifiers maintain precision better than the CNN-based models for 0.52 < z < 1. The authors conclude that simple morphological cuts remain competitive for high-purity samples and that new statistics are required for post-mergers.

Significance. If the reported precisions and robustness hold, the work supplies practical, publicly usable linear cuts that recover purity comparable to contemporary CNNs and transformers while being far cheaper to apply and more stable under domain shift. The explicit extension of the MB24 merger challenge, the shared train/test split, the MCMC optimisation with validation regularisation, and the open segmentation code are concrete strengths that make the comparison reproducible and the criteria immediately usable for large surveys. The clear demonstration that G, M20 and C are blind to post-mergers is itself a useful negative result that motivates new morphological statistics.

major comments (3)
  1. §5.3–5.4 and Table 2: the HSC-SSP evaluation yields precisions of 98–100 % while recall stays <50 %. The text correctly attributes this to the addition of ~1000 “clear non-mergers,” yet still reports these numbers as headline performance and includes them in the robustness table. Because the visual ground truth is demonstrably biased, the HSC-SSP metrics cannot support claims of real-data performance. Either re-evaluate on a less biased visual sample or move the HSC-SSP numbers to a clearly labelled “biased-sample” appendix and base all robustness conclusions solely on Horizon-AGN.
  2. §3.4, Eq. (13) and Table 1: the likelihood is the reciprocal of accuracy (exponentiated) and the priors/initial conditions are set by visual inspection of the “turnover region” (Appendix B). Accuracy is a poor objective for an unbalanced morphological distribution; maximising precision (or F1) subject to a minimum recall would better match the paper’s own emphasis on purity. The visual prior selection should be replaced by, or at least compared against, a fully data-driven prior (e.g., from a logistic-regression or SVM fit) so that the final slopes and intercepts are not sensitive to the eye-fit step.
  3. §5.2 and Figs. 12–14: the central claim that the classifiers “select only pre-mergers” is well supported, but the paper never quantifies how much of the low recall is simply the fraction of the merger sample that is post/ongoing (0.1 Gyr before to 0.3 Gyr after). A simple breakdown of recall by the four time bins already shown in Fig. 14, together with the fraction of the training set that falls into each bin, would make the pre-merger-only conclusion quantitative rather than qualitative and would strengthen the comparison with ML methods that can recover post-mergers.
minor comments (5)
  1. Throughout: “IlustrisTNG”, “Simualtions”, “Petrossian”, “liens”, “form this work” and similar typos should be corrected.
  2. Fig. 3 caption and §4.1: the confusion-matrix cells are described as containing “precision, recall, and number,” but the vertical normalisation is actually the fraction of predicted labels (precision) while the horizontal is the fraction of true labels (recall). Clarify the wording so that readers do not confuse the two.
  3. Eq. (16): the robustness metric ρ is defined only for precision. A parallel definition for recall (or F1) would make the redshift-dependence discussion more complete.
  4. Appendix A: the morphological histograms for Horizon-AGN show a small tail of extremely low M20; a brief comment on whether these objects are real or segmentation artefacts would be useful.
  5. Table 2: bolding the highest value in each column is helpful, but the morphological rows should also be highlighted when they fall inside the ML range so that the “consistent with” claim is immediately visible.

Circularity Check

0 steps flagged

No significant circularity: MCMC cuts are fitted on training data and evaluated on held-out/transfer sets; claimed precisions and robustness are independent measurements, not forced by construction.

full rationale

The paper optimizes linear separators in G–M20 and G–C space via MCMC on the MB24 IllustrisTNG training split (likelihood = inverse accuracy, with validation-set rejection of overfits and explicit priors), then reports precision/recall/accuracy/F1 on the held-out IllustrisTNG test split, Horizon-AGN mocks, and HSC-SSP observations. The reported numbers (e.g., 69.5 % / 72.3 % precision) are therefore ordinary out-of-sample metrics, not algebraic identities of the fitted parameters. The robustness statistic ρ is a simple post-hoc ratio of those precisions and does not enter the optimization. Comparison to the six MB24 ML models uses the identical train/test split by design for fairness; shared merger-tree labels are a common experimental condition, not a self-definitional reduction of the morphological result. Literature cuts (Lotz et al.) are re-evaluated on the same data and shown to be slightly inferior, confirming the new cuts are not a renaming. No uniqueness theorem, ansatz smuggled via self-citation, or fitted constant re-labeled as a prediction appears in the derivation chain. The work is therefore self-contained against its external benchmarks.

Axiom & Free-Parameter Ledger

3 free parameters · 3 axioms · 0 invented entities

Central claim rests on standard morphological definitions, the MB24 merger-tree labels, and two free parameters per linear cut fitted by MCMC. No new physical entities are postulated.

free parameters (3)
  • G-M20 slope and intercept = m=-0.267, c=0.143
    Linear separator G = m M20 + c fitted by MCMC on IllustrisTNG training set; reported as m = -0.267 ± 0.081, c = 0.143 ± 0.012.
  • G-C slope and intercept = m=0.162, c=-0.149
    Linear separator G = m C + c fitted by MCMC; reported as m = 0.162 ± 0.048, c = -0.149 ± 0.12.
  • segmentation deblend distance threshold = 32 pixels
    Adjacent segments whose brightest pixels are <32 pixels apart are merged; chosen by hand to correct over-segmentation of mergers.
axioms (3)
  • domain assumption Major-merger labels from SubLink/AdaptaHOP trees (mass ratio >1:4, -0.8 to +0.3 Gyr around coalescence) are the ground truth for both morphological and ML classifiers.
    Section 2.1; inherited from MB24 and used for all precision/recall calculations.
  • domain assumption Gini, M20 and concentration computed by statmorph on the central segmented source are sufficient statistics for the morphological comparison.
    Section 3.2; standard Lotz/Conselice definitions adopted without modification.
  • domain assumption Mock HSC-SSP images (PSF convolution + Poisson noise + real backgrounds) are realistic enough for transfer tests to real HSC data.
    Section 2.1.3; same mocks used by MB24.

pith-pipeline@v1.1.0-grok45 · 29583 in / 2645 out tokens · 25267 ms · 2026-07-13T04:39:47.208304+00:00 · methodology

0 comments
read the original abstract

Aims. Non-parametric morphological statistics can be used for efficient classification of galaxy mergers. This work aims to compare the performance of morphological merger classifiers to state-of-the-art machine learning (ML) models. A secondary aim is to produce updated criteria for mergers based on non-parametric morphological statistics. Methods. The Gini coefficient (G), $M_{20}$ statistic, and concentration ($C$) were calculated for mock Hyper Suprime-Cam Subaru Strategic Program (HSC-SSP) images based on the IllustrisTNG and Horizon-AGN simulations, and observations from HSC-SSP. The IllustrisTNG images were used to find the line which best separates mergers and non-mergers in 2D morphological space with a Markov Chain Monte-Carlo (MCMC) method. Results. Based on the MCMC results, we classified galaxies with $G>(-0.267\pm0.081)M_{20}+(0.143\pm0.012)$ or $G>(0.162\pm0.048)C-(0.149\pm0.12)$ as mergers, these criteria had precisions of 69.5\% and 72.3\% respectively when applied to previously unseen IllustrisTNG mock HSC-SSP images. The precisions of the morphological classifications are consistent with state-of-the-art ML methods. The morphological classifiers were found to be effective at selecting only pre-mergers; post-merger galaxies are indistinguishable from non-mergers in terms of their $G$, $M_{20}$, and $C$ values. Morphological classifiers displayed a similar robustness to new data to ML methods up to a redshift of $\sim0.52$ and maintained robustness better than ML methods based on convolutional neural networks in the redshift range $0.52<z<1$. Conclusions. This work presents updated morphological classifiers which achieve similar precisions to ML based merger classifiers with a high robustness to new data. New morphological statistics are needed to identify the features of post-merger galaxies.

Figures

Figures reproduced from arXiv: 2607.09209 by Aidan P. Cotter, Alejandro Guzm\'an-Ortega, Berta Margalef-Bentabol, Subhrata Dey, Vicente Rodriguez-Gomez, William J. Pearson.

Figure 1
Figure 1. Figure 1: Example of the segmentation of a mock image of a merger from the IllustrisTNG simulation. Left most panel shows the [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: The distribution of the training data in [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Confusion matrices displaying the performance of [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Distribution of merger subclasses in the testing data in [PITH_FULL_IMAGE:figures/full_fig_p008_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Confusion matrices displaying the performance of the [PITH_FULL_IMAGE:figures/full_fig_p009_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Performance of the merger classifiers with binned redshift in terms of precision (left) and recall (right). Red stars represent [PITH_FULL_IMAGE:figures/full_fig_p010_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Performance of the merger classifiers with binned stellar mass in terms of precision (left) and recall (right). The symbols and [PITH_FULL_IMAGE:figures/full_fig_p010_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Confusion matrices displaying the performance of [PITH_FULL_IMAGE:figures/full_fig_p012_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: 2D histogram of the G − M20 (left) and G − C (right) distributions of the training sample from MB24. Top panels include non-merging galaxies and the bottom panels include merging galaxies. The uniformly sized bins are displayed by rectangles scaled logarithmically by their population, the colouring shows the median stellar mass in each bin from low (in purple) to high (in red). The optimised criterion is i… view at source ↗
Figure 10
Figure 10. Figure 10: 2D histogram of the G − M20 (left) and G − C (right) distributions of the mergers in the training sample from MB24. The uniformly sized bins are displayed by rectangles scaled logarithmically by their population, the colouring shows the median time before the next merger in each bin from low (in purple) to high (in red). The black lines represent the same criteria as in [PITH_FULL_IMAGE:figures/full_fig_… view at source ↗
Figure 11
Figure 11. Figure 11: 2D histogram of the G − M20 (left) and G − C (right) distributions of the training sample from MB24. Top panels include non-merging galaxies and the bottom panels include merging galaxies. The uniformly sized bins are displayed by rectangles scaled logarithmically by their population, the colouring shows the median stellar mass in each bin from low (in purple) to high (in red). The black lines represent t… view at source ↗
Figure 12
Figure 12. Figure 12: Contours showing the distribution of pre-mergers (in red) and post-mergers (in blue) from the training sample in the [PITH_FULL_IMAGE:figures/full_fig_p014_12.png] view at source ↗
Figure 13
Figure 13. Figure 13: Distribution of the ongoing mergers (bottom panels) and post mergers (top panels) from the training sample in the [PITH_FULL_IMAGE:figures/full_fig_p015_13.png] view at source ↗
Figure 15
Figure 15. Figure 15: The robustness of the G-M20 and G-C classifiers (in red and blue respectively) compared to the 6 ML classifiers from MB24. Robustness is calculated from the precision on the testing data and the Horizon-AGN data. Acknowledgements. The authors would like to thank Krzysiek Lisiecki for their software contribution. A.C., W.J.P., and S.D. have been supported by the Polish National Science Center project UMO-2… view at source ↗

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

106 extracted references · 29 canonical work pages · 26 internal anchors

  1. [1]

    arXiv e-prints , keywords =

    Detection of the 2175 A UV Bump at z>7: Evidence for Rapid Dust Evolution in a Merging Reionisation-Era Galaxy. arXiv e-prints , keywords =. doi:10.48550/arXiv.2502.21119 , archivePrefix =. 2502.21119 , primaryClass =

    Pith/arXiv arXiv doi:10.48550/arxiv.2502.21119

  2. [2]

    The major merger-active galactic nucleus connection up to the cosmic noon

    The major merger-active galactic nucleus connection up to the cosmic noon. arXiv e-prints , keywords =. doi:10.48550/arXiv.2602.19997 , archivePrefix =. 2602.19997 , primaryClass =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2602.19997

  3. [3]

    , keywords =

    Galaxy morphologies at cosmic noon with JWST: A foundation for exploring gas transport with bars and spiral arms. , keywords =. doi:10.1051/0004-6361/202554725 , archivePrefix =. 2503.21738 , primaryClass =

    Pith/arXiv arXiv doi:10.1051/0004-6361/202554725

  4. [4]

    A New Approach to Galaxy Morphology. I. Analysis of the Sloan Digital Sky Survey Early Data Release. , keywords =. doi:10.1086/373919 , archivePrefix =. astro-ph/0301239 , primaryClass =

    Pith/arXiv arXiv doi:10.1086/373919

  5. [5]

    Evidence for Multiple Mergers among Ultraluminous IR Galaxies (ULIRGs): Remnants of Compact Groups?

    Evidence for Multiple Mergers among Ultraluminous Infrared Galaxies: Remnants of Compact Groups?. , keywords =. doi:10.1086/312461 , archivePrefix =. astro-ph/9912151 , primaryClass =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.1086/312461

  6. [6]

    , keywords =

    Galaxy Zoo: CANDELS barred discs and bar fractions. , keywords =. doi:10.1093/mnras/stu1817 , archivePrefix =. 1409.1214 , primaryClass =

    Pith/arXiv arXiv doi:10.1093/mnras/stu1817

  7. [7]

    Simulations of binary galaxy mergers and the link with fast rotators, slow rotators and kinematically distinct cores

    The ATLAS ^ 3D project - VI. Simulations of binary galaxy mergers and the link with fast rotators, slow rotators and kinematically distinct cores. , keywords =. doi:10.1111/j.1365-2966.2011.19113.x , archivePrefix =. 1105.4076 , primaryClass =

    doi:10.1111/j.1365-2966.2011.19113.x 2011

  8. [8]

    The unorthodox evolution of major merger remnants into star-forming spiral galaxies

    The unorthodox evolution of major merger remnants into star-forming spiral galaxies. , keywords =. doi:10.1093/mnras/stx1516 , archivePrefix =. 1610.03850 , primaryClass =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.1093/mnras/stx1516

  9. [9]

    Galaxy Zoo JWST: Up to 75% of discs are featureless at $3<z<7$

    Galaxy Zoo JWST: up to 75 per cent of discs are featureless at 3 < z < 7. , keywords =. doi:10.1093/mnras/staf506 , archivePrefix =. 2503.21869 , primaryClass =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.1093/mnras/staf506

  10. [10]

    , year = 1996, month = jan, volume =

    Luminous Infrared Galaxies. , year = 1996, month = jan, volume =. doi:10.1146/annurev.astro.34.1.749 , adsurl =

    doi:10.1146/annurev.astro.34.1.749 1996

  11. [11]

    , keywords =

    Automated distant galaxy merger classifications from Space Telescope images using the Illustris simulation. , keywords =. doi:10.1093/mnras/stz1059 , archivePrefix =. 1809.02136 , primaryClass =

    Pith/arXiv arXiv doi:10.1093/mnras/stz1059

  12. [12]

    Quasi-Stellar Objects, Ultraluminous IR Galaxies, and Mergers

    Quasi-Stellar Objects, Ultraluminous Infrared Galaxies, and Mergers. , keywords =. doi:10.1086/321520 , archivePrefix =. astro-ph/0103332 , primaryClass =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.1086/321520

  13. [13]

    , keywords =

    Properties of galaxies reproduced by a hydrodynamic simulation. , keywords =. doi:10.1038/nature13316 , archivePrefix =. 1405.1418 , primaryClass =

    Pith/arXiv arXiv doi:10.1038/nature13316

  14. [14]

    , keywords =

    E pur si muove: Galilean-invariant cosmological hydrodynamical simulations on a moving mesh. , keywords =. doi:10.1111/j.1365-2966.2009.15715.x , archivePrefix =. 0901.4107 , primaryClass =

    doi:10.1111/j.1365-2966.2009.15715.x 2009

  15. [15]

    , keywords =

    Star Formation in Galaxies Along the Hubble Sequence. , keywords =. doi:10.1146/annurev.astro.36.1.189 , archivePrefix =. astro-ph/9807187 , primaryClass =

    Pith/arXiv arXiv doi:10.1146/annurev.astro.36.1.189

  16. [16]

    , keywords =

    Galactic Stellar and Substellar Initial Mass Function. , keywords =. doi:10.1086/376392 , archivePrefix =. astro-ph/0304382 , primaryClass =

    Pith/arXiv arXiv doi:10.1086/376392

  17. [17]

    , keywords =

    The evolution of large-scale structure in a universe dominated by cold dark matter. , keywords =. doi:10.1086/163168 , adsurl =

    doi:10.1086/163168

  18. [18]

    Results at z=0

    Populating a cluster of galaxies - I. Results at z=0. , keywords =. doi:10.1046/j.1365-8711.2001.04912.x , archivePrefix =. astro-ph/0012055 , primaryClass =

    doi:10.1046/j.1365-8711.2001.04912.x 2001

  19. [19]

    The Journal of Open Source Software , year = 2016, month = oct, volume =

    SEP: Source Extractor as a library. The Journal of Open Source Software , year = 2016, month = oct, volume =. doi:10.21105/joss.00058 , adsurl =

    doi:10.21105/joss.00058 2016

  20. [20]

    Galaxy pairs in the Sloan Digital Sky Survey -- XIV. Galaxy mergers do not lie on the Fundamental Metallicity Relation

    Galaxy pairs in the Sloan Digital Sky Survey - XIV. Galaxy mergers do not lie on the fundamental metallicity relation. , keywords =. doi:10.1093/mnras/staa1025 , archivePrefix =. 2004.06121 , primaryClass =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.1093/mnras/staa1025 2004

  21. [21]

    Planck 2015 results. XIII. Cosmological parameters. , keywords =. doi:10.1051/0004-6361/201525830 , archivePrefix =. 1502.01589 , primaryClass =

    Pith/arXiv arXiv doi:10.1051/0004-6361/201525830 2015

  22. [22]

    , keywords =

    Substructures in hydrodynamical cluster simulations. , keywords =. doi:10.1111/j.1365-2966.2009.15034.x , archivePrefix =. 0808.3401 , primaryClass =

    doi:10.1111/j.1365-2966.2009.15034.x 2009

  23. [23]

    , keywords =

    Introducing the Illustris Project: simulating the coevolution of dark and visible matter in the Universe. , keywords =. doi:10.1093/mnras/stu1536 , archivePrefix =. 1405.2921 , primaryClass =

    Pith/arXiv arXiv doi:10.1093/mnras/stu1536

  24. [24]

    , keywords =

    Introducing the Illustris project: the evolution of galaxy populations across cosmic time. , keywords =. doi:10.1093/mnras/stu1654 , archivePrefix =. 1405.3749 , primaryClass =

    Pith/arXiv arXiv doi:10.1093/mnras/stu1654

  25. [25]

    , keywords =

    The Illustris simulation: the evolving population of black holes across cosmic time. , keywords =. doi:10.1093/mnras/stv1340 , archivePrefix =. 1408.6842 , primaryClass =

    Pith/arXiv arXiv doi:10.1093/mnras/stv1340

  26. [26]

    Galaxy Interactions Trigger Rapid Black Hole Growth: an unprecedented view from the Hyper Suprime-Cam Survey

    Galaxy interactions trigger rapid black hole growth: An unprecedented view from the Hyper Suprime-Cam survey. , keywords =. doi:10.1093/pasj/psx135 , archivePrefix =. 1706.07436 , primaryClass =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.1093/pasj/psx135

  27. [27]

    The Open Journal of Astrophysics , keywords =

    RMS asymmetry: a robust metric of galaxy shapes in images with varied depth and resolution. The Open Journal of Astrophysics , keywords =. doi:10.33232/001c.123524 , archivePrefix =. 2404.05792 , primaryClass =

    Pith/arXiv arXiv doi:10.33232/001c.123524

  28. [28]

    Galaxy merger challenge: A comparison study between machine learning-based detection methods

    Galaxy merger challenge: A comparison study between machine learning-based detection methods. , keywords =. doi:10.1051/0004-6361/202348239 , archivePrefix =. 2403.15118 , primaryClass =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.1051/0004-6361/202348239

  29. [29]

    , keywords =

    emcee: The MCMC Hammer. , keywords =. doi:10.1086/670067 , archivePrefix =. 1202.3665 , primaryClass =

    Pith/arXiv arXiv doi:10.1086/670067

  30. [30]

    , keywords =

    The Hyper Suprime-Cam software pipeline. , keywords =. doi:10.1093/pasj/psx080 , archivePrefix =. 1705.06766 , primaryClass =

    Pith/arXiv arXiv doi:10.1093/pasj/psx080

  31. [31]

    , keywords =

    First results from the IllustrisTNG simulations: a tale of two elements - chemical evolution of magnesium and europium. , keywords =. doi:10.1093/mnras/sty618 , archivePrefix =. 1707.03401 , primaryClass =

    Pith/arXiv arXiv doi:10.1093/mnras/sty618

  32. [32]

    , keywords =

    First results from the IllustrisTNG simulations: the stellar mass content of groups and clusters of galaxies. , keywords =. doi:10.1093/mnras/stx3112 , archivePrefix =. 1707.03406 , primaryClass =

    Pith/arXiv arXiv doi:10.1093/mnras/stx3112

  33. [33]

    , keywords =

    First results from the IllustrisTNG simulations: matter and galaxy clustering. , keywords =. doi:10.1093/mnras/stx3304 , archivePrefix =. 1707.03397 , primaryClass =

    Pith/arXiv arXiv doi:10.1093/mnras/stx3304

  34. [34]

    , keywords =

    First results from the IllustrisTNG simulations: the galaxy colour bimodality. , keywords =. doi:10.1093/mnras/stx3040 , archivePrefix =. 1707.03395 , primaryClass =

    Pith/arXiv arXiv doi:10.1093/mnras/stx3040

  35. [35]

    , keywords =

    First results from the IllustrisTNG simulations: radio haloes and magnetic fields. , keywords =. doi:10.1093/mnras/sty2206 , archivePrefix =. 1707.03396 , primaryClass =

    Pith/arXiv arXiv doi:10.1093/mnras/sty2206

  36. [36]

    ERGO-ML I: Inferring the assembly histories of IllustrisTNG galaxies from integral observable properties via invertible neural networks

    ERGO-ML I: inferring the assembly histories of IllustrisTNG galaxies from integral observable properties via invertible neural networks. , keywords =. doi:10.1093/mnras/stac3295 , archivePrefix =. 2202.06967 , primaryClass =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.1093/mnras/stac3295

  37. [37]

    Communications in Applied Mathematics and Computational Science , keywords =

    Ensemble samplers with affine invariance. Communications in Applied Mathematics and Computational Science , keywords =. doi:10.2140/camcos.2010.5.65 , adsurl =

    doi:10.2140/camcos.2010.5.65 2010

  38. [38]

    The Symmetry, Color, and Morphology of Galaxies

    The Symmetry, Color, and Morphology of Galaxies. , keywords =. doi:10.1086/134004 , archivePrefix =. astro-ph/9710234 , primaryClass =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.1086/134004

  39. [39]

    , keywords =

    The Asymmetry of Galaxies: Physical Morphology for Nearby and High-Redshift Galaxies. , keywords =. doi:10.1086/308300 , archivePrefix =. astro-ph/9907399 , primaryClass =

    Pith/arXiv arXiv doi:10.1086/308300

  40. [40]

    , keywords =

    Core condensation in heavy halos: a two-stage theory for galaxy formation and clustering. , keywords =. doi:10.1093/mnras/183.3.341 , adsurl =

    doi:10.1093/mnras/183.3.341

  41. [41]

    , keywords =

    Galaxy Formation through Hierarchical Clustering. , keywords =. doi:10.1086/170483 , adsurl =

    doi:10.1086/170483

  42. [42]

    Convolutional neural network identification of galaxy post-mergers in UNIONS using IllustrisTNG

    Convolutional neural network identification of galaxy post-mergers in UNIONS using IllustrisTNG. , keywords =. doi:10.1093/mnras/stab806 , archivePrefix =. 2103.09367 , primaryClass =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.1093/mnras/stab806

  43. [43]

    , keywords =

    Dancing in the dark: galactic properties trace spin swings along the cosmic web. , keywords =. doi:10.1093/mnras/stu1227 , archivePrefix =. 1402.1165 , primaryClass =

    Pith/arXiv arXiv doi:10.1093/mnras/stu1227

  44. [44]

    , keywords =

    Surface Brightness and Evolution of Galaxies. , keywords =. doi:10.1086/18230110.1086/182253 , adsurl =

    doi:10.1086/18230110.1086/182253

  45. [45]

    Classifying merger stages with adaptive deep learning and cosmological hydrodynamical simulations

    Classifying merger stages with adaptive deep learning and cosmological hydrodynamical simulations. , keywords =. doi:10.1051/0004-6361/202452659 , archivePrefix =. 2502.06254 , primaryClass =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.1051/0004-6361/202452659

  46. [46]

    Galaxy Mergers in UNIONS -- I: A Simulation-driven Hybrid Deep Learning Ensemble for Pure Galaxy Merger Classification

    Galaxy mergers in UNIONS - I. A simulation-driven hybrid deep learning ensemble for pure galaxy merger classification. , keywords =. doi:10.1093/mnras/stae1885 , archivePrefix =. 2407.18396 , primaryClass =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.1093/mnras/stae1885

  47. [47]

    Planck 2013 results. XVI. Cosmological parameters. , keywords =. doi:10.1051/0004-6361/201321591 , archivePrefix =. 1303.5076 , primaryClass =

    Pith/arXiv arXiv doi:10.1051/0004-6361/201321591 2013

  48. [48]

    , keywords =

    The Hyper Suprime-Cam SSP Survey: Overview and survey design. , keywords =. doi:10.1093/pasj/psx066 , archivePrefix =. 1704.05858 , primaryClass =

    Pith/arXiv arXiv doi:10.1093/pasj/psx066

  49. [49]

    , keywords =

    The origin and implications of dark matter anisotropic cosmic infall on -0.5ex _ * haloes. , keywords =. doi:10.1111/j.1365-2966.2004.07883.x , archivePrefix =. astro-ph/0402405 , primaryClass =

    doi:10.1111/j.1365-2966.2004.07883.x 2004

  50. [50]

    Towards improving galaxy formation models using subhaloes

    Building merger trees from cosmological N-body simulations. Towards improving galaxy formation models using subhaloes. , keywords =. doi:10.1051/0004-6361/200911787 , archivePrefix =. 0902.0679 , primaryClass =

    Pith/arXiv arXiv doi:10.1051/0004-6361/200911787

  51. [51]

    , year = 1955, month = jan, volume =

    The Luminosity Function and Stellar Evolution. , year = 1955, month = jan, volume =. doi:10.1086/145971 , adsurl =

    doi:10.1086/145971 1955

  52. [52]

    Radiative Transfer in a Clumpy Universe. II. The Ultraviolet Extragalactic Background. , keywords =. doi:10.1086/177035 , archivePrefix =. astro-ph/9509093 , primaryClass =

    Pith/arXiv arXiv doi:10.1086/177035

  53. [53]

    , keywords =

    Self-regulated growth of supermassive black holes by a dual jet-heating active galactic nucleus feedback mechanism: methods, tests and implications for cosmological simulations. , keywords =. doi:10.1111/j.1365-2966.2011.20236.x , archivePrefix =. 1108.0110 , primaryClass =

    doi:10.1111/j.1365-2966.2011.20236.x 2011

  54. [54]

    , keywords =

    Cooling Functions for Low-Density Astrophysical Plasmas. , keywords =. doi:10.1086/191823 , adsurl =

    doi:10.1086/191823

  55. [55]

    Initial Conditions for Large Cosmological Simulations

    Initial Conditions For Large Cosmological Simulations. , keywords =. doi:10.1086/590370 , archivePrefix =. 0804.3536 , primaryClass =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.1086/590370

  56. [56]

    , keywords =

    Seven-year Wilkinson Microwave Anisotropy Probe (WMAP) Observations: Cosmological Interpretation. , keywords =. doi:10.1088/0067-0049/192/2/18 , archivePrefix =. 1001.4538 , primaryClass =

    Pith/arXiv arXiv doi:10.1088/0067-0049/192/2/18

  57. [57]

    A new high resolution code called RAMSES

    Cosmological hydrodynamics with adaptive mesh refinement. A new high resolution code called RAMSES. , keywords =. doi:10.1051/0004-6361:20011817 , archivePrefix =. astro-ph/0111367 , primaryClass =

    Pith/arXiv arXiv doi:10.1051/0004-6361:20011817

  58. [58]

    , keywords =

    Using transfer learning to detect galaxy mergers. , keywords =. doi:10.1093/mnras/sty1398 , archivePrefix =. 1805.10289 , primaryClass =

    Pith/arXiv arXiv doi:10.1093/mnras/sty1398

  59. [59]

    , keywords =

    The nearly universal merger rate of dark matter haloes in CDM cosmology. , keywords =. doi:10.1111/j.1365-2966.2008.13075.x , archivePrefix =. 0710.4567 , primaryClass =

    doi:10.1111/j.1365-2966.2008.13075.x 2008

  60. [60]

    Morphological signatures of mergers in the TNG50 simulation and the Kilo-Degree Survey: the merger fraction from dwarfs to Milky Way-like galaxies

    Morphological signatures of mergers in the TNG50 simulation and the Kilo-Degree Survey: the merger fraction from dwarfs to Milky Way-like galaxies. , keywords =. doi:10.1093/mnras/stac3334 , archivePrefix =. 2211.05785 , primaryClass =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.1093/mnras/stac3334

  61. [61]

    North Ecliptic Pole merging galaxy catalogue

    North Ecliptic Pole merging galaxy catalogue. , keywords =. doi:10.1051/0004-6361/202141013 , archivePrefix =. 2202.10780 , primaryClass =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.1051/0004-6361/202141013

  62. [62]

    , keywords =

    LSST: From Science Drivers to Reference Design and Anticipated Data Products. , keywords =. doi:10.3847/1538-4357/ab042c , archivePrefix =. 0805.2366 , primaryClass =

    Pith/arXiv arXiv doi:10.3847/1538-4357/ab042c

  63. [63]

    Are Adversarial Robustness and Common Perturbation Robustness Independent Attributes ?

    Are Adversarial Robustness and Common Perturbation Robustness Independent Attributes ?. arXiv e-prints , keywords =. doi:10.48550/arXiv.1909.02436 , archivePrefix =. 1909.02436 , primaryClass =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.1909.02436 1909

  64. [64]

    III The triggering of active galactic nuclei peaks immediately after coalescence

    Galaxy evolution in the post-merger regime. III The triggering of active galactic nuclei peaks immediately after coalescence. The Open Journal of Astrophysics , keywords =. doi:10.33232/001c.129235 , archivePrefix =. 2412.02804 , primaryClass =

    Pith/arXiv arXiv doi:10.33232/001c.129235

  65. [65]

    , keywords =

    Active galactic nuclei and galaxy interactions. , keywords =. doi:10.1111/j.1365-2966.2007.11367.x , archivePrefix =. astro-ph/0701192 , primaryClass =

    doi:10.1111/j.1365-2966.2007.11367.x 2007

  66. [66]

    Overview of the Euclid mission

    Euclid: I. Overview of the Euclid mission. , keywords =. doi:10.1051/0004-6361/202450810 , archivePrefix =. 2405.13491 , primaryClass =

    doi:10.1051/0004-6361/202450810

  67. [67]

    , keywords =

    Galaxy Zoo: the properties of merging galaxies in the nearby Universe - local environments, colours, masses, star formation rates and AGN activity. , keywords =. doi:10.1111/j.1365-2966.2009.15786.x , archivePrefix =. 0903.5057 , primaryClass =

    doi:10.1111/j.1365-2966.2009.15786.x 2009

  68. [68]

    , keywords =

    Observational evidence for AGN feedback in early-type galaxies. , keywords =. doi:10.1111/j.1365-2966.2007.12487.x , archivePrefix =. 0709.3015 , primaryClass =

    doi:10.1111/j.1365-2966.2007.12487.x 2007

  69. [69]

    , keywords =

    Effect of galaxy mergers on star-formation rates. , keywords =. doi:10.1051/0004-6361/201936337 , archivePrefix =. 1908.10115 , primaryClass =

    Pith/arXiv arXiv doi:10.1051/0004-6361/201936337 1908

  70. [70]

    , keywords =

    Shock-induced star formation in a model of the Mice. , keywords =. doi:10.1111/j.1365-2966.2004.07725.x , archivePrefix =. astro-ph/0402248 , primaryClass =

    doi:10.1111/j.1365-2966.2004.07725.x 2004

  71. [71]

    Toward First-Principle Simulations of Galaxy Formation: II. Shock-Induced Starburst at a Collision Interface During the First Encounter of Interacting Galaxies

    Toward First-Principle Simulations of Galaxy Formation: II. Shock-Induced Starburst at a Collision Interface during the First Encounter of Interacting Galaxies. , keywords =. doi:10.1093/pasj/61.3.481 , archivePrefix =. 0805.0167 , primaryClass =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.1093/pasj/61.3.481

  72. [72]

    , keywords =

    The optical morphologies of galaxies in the IllustrisTNG simulation: a comparison to Pan-STARRS observations. , keywords =. doi:10.1093/mnras/sty3345 , archivePrefix =. 1809.08239 , primaryClass =

    Pith/arXiv arXiv doi:10.1093/mnras/sty3345

  73. [73]

    , keywords =

    The Astropy Project: Sustaining and Growing a Community-oriented Open-source Project and the Latest Major Release (v5.0) of the Core Package. , keywords =. doi:10.3847/1538-4357/ac7c74 , archivePrefix =. 2206.14220 , primaryClass =

    Pith/arXiv arXiv doi:10.3847/1538-4357/ac7c74

  74. [74]

    and Haberland, Matt and Reddy, Tyler and Cournapeau, David and Burovski, Evgeni and Peterson, Pearu and Weckesser, Warren and Bright, Jonathan and

    Virtanen, Pauli and Gommers, Ralf and Oliphant, Travis E. and Haberland, Matt and Reddy, Tyler and Cournapeau, David and Burovski, Evgeni and Peterson, Pearu and Weckesser, Warren and Bright, Jonathan and. Nature Methods , year =

  75. [75]

    , keywords =

    The Astropy Project: Building an Open-science Project and Status of the v2.0 Core Package. , keywords =. doi:10.3847/1538-3881/aabc4f , archivePrefix =. 1801.02634 , primaryClass =

    Pith/arXiv arXiv doi:10.3847/1538-3881/aabc4f

  76. [76]

    , keywords =

    Astropy: A community Python package for astronomy. , keywords =. doi:10.1051/0004-6361/201322068 , archivePrefix =. 1307.6212 , primaryClass =

    Pith/arXiv arXiv doi:10.1051/0004-6361/201322068

  77. [77]

    , keywords =

    Constraining the major merger history of z -0.5ex 3-9 galaxies using JADES: dominant in situ star formation. , keywords =. doi:10.1093/mnras/staf813 , archivePrefix =. 2502.01721 , primaryClass =

    Pith/arXiv arXiv doi:10.1093/mnras/staf813

  78. [78]

    , keywords =

    Dynamically Close Galaxy Pairs and Merger Rate Evolution in the CNOC2 Redshift Survey. , keywords =. doi:10.1086/324543 , archivePrefix =. astro-ph/0109428 , primaryClass =

    Pith/arXiv arXiv doi:10.1086/324543

  79. [79]

    A direct measurement of galaxy major and minor merger rates and stellar mass accretion histories at $z < 3$ using galaxy pairs in the REFINE survey

    A Direct Measurement of Galaxy Major and Minor Merger Rates and Stellar Mass Accretion Histories at Z < 3 Using Galaxy Pairs in the REFINE Survey. , keywords =. doi:10.3847/1538-4357/ac9b1a , archivePrefix =. 2207.03984 , primaryClass =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.3847/1538-4357/ac9b1a

  80. [80]

    The ALHAMBRA survey: Accurate merger fractions by PDF analysis of photometric close pairs

    The ALHAMBRA survey: accurate merger fractions derived by PDF analysis of photometrically close pairs. , keywords =. doi:10.1051/0004-6361/201424913 , archivePrefix =. 1409.1142 , primaryClass =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.1051/0004-6361/201424913

Showing first 80 references.