Pith. sign in

REVIEW 4 major objections 7 minor 62 references

Selection of Dwarf Galaxies Hosting AGNs: A Measure of Bias and Contamination using Unsupervised Machine Learning Techniques

T0 review · 4 major / 7 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read Unsupervised machine learning on 30,344 dwarf galaxy SEDs reveals that each AGN detection method selects a distinct host population, with WISE-selected AGNs splitting into a starburst-like clump and a redder, AGN-like clump.

desk verdict A useful SOM-based map of dwarf AGN selection, but the WISE clump interpretation rests on an unvalidated line-ratio proxy. read the letter →

arxiv 2505.16509 v1 pith:YNKUYK7L submitted 2025-05-22 astro-ph.GA

Sogol Sanjaripour , Archana Aravindan , Gabriela Canalizo , Shoubaneh Hemmati , Bahram Mobasher , Alison L. Coil , Barry C. Barish This is my paper · ORCID
classification astro-ph.GA
keywords AGNselectiondwarfgalaxiesself-organizingmapsspectralenergydistributionsWISEmid-infraredbiasstarburstcontaminationunsupervisedmachinelearning
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 shows that the way we find active black holes in dwarf galaxies is heavily shaped by the type of galaxy each detection method can see. Using a self-organizing map (SOM, a machine-learning grid that places similar galaxy colors next to each other) trained on the photometric colors of 30,344 dwarf galaxies, the authors put 438 previously known dwarf AGNs on a single map of spectral energy distributions and find that AGNs selected by different diagnostics cluster in different, partially overlapping regions. BPT-selected AGNs (found through optical emission-line ratios) sit in higher-mass hosts, X-ray and variability-selected AGNs spread widely, and WISE-selected AGNs (found through mid-infrared colors) concentrate in low-mass hosts where they form two clumps: one bluer and starburst-like, the other redder with AGN-like spectral signatures. If true, this gives a concrete measure of the selection bias and contamination in each technique, and the WISE clump split offers a way to clean starburst contaminants out of infrared-selected dwarf AGN samples.

What carries the argument

The load-bearing object is a $51\times 51$ Self-Organizing Map (SOM), an unsupervised neural network that compresses the seven-dimensional color space built from nine photometric bands (FUV, NUV, $u$, $g$, $r$, $i$, $z$, W1, W2) into a two-dimensional grid while preserving which spectral energy distributions resemble which. It is trained on 30,344 dwarf galaxies from the NSA catalog with $z<0.055$ and $M_*<10^{9.5}\,M_\odot$, and the 438 known AGNs are overlaid cell by cell. The argument that the two WISE clumps differ physically is carried by two proxy maps: $[\mathrm{O\,II}]/z$-band flux for star-formation activity and $[\mathrm{O\,III}]/([\mathrm{O\,II}]\times z)$ for AGN luminosity relative to stellar light, supplemented by visual inspection of SDSS images and color–mass diagrams for the clump galaxies.

What would settle it

Conduct a spectroscopic follow-up of WISE-selected dwarf AGNs from both SOM clumps and measure dust-corrected BPT line ratios or coronal emission lines. If the bluer clump shows the same AGN fraction as the redder clump—or if the two clumps have equal X-ray detection rates—the claimed separation into starburst contaminants and true AGN hosts fails.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central discovery is that AGN diagnostics are not interchangeable windows onto the same dwarf-galaxy population: each method samples a distinct and only partially overlapping region of the spectral energy distribution manifold. The strongest new result is that WISE-selected dwarf AGNs appear in two separated clumps on the map, with the right clump made of bluer, actively star-forming galaxies and the left clump of redder galaxies whose spectral proxies point to AGN activity; the left clump overlaps heavily with AGNs found by BPT, broad-line, X-ray, [O I], [S II], and He II selection, while the right clump overlaps with regions of high $[\mathrm{O\,II}]/z$ star-formation tracers. The paper interprets this as the SOM separating true AGN hosts from starburst contaminants in WISE selection, and it further finds that a subset of AGNs in low-mass galaxies occupy cells with high $[\mathrm{O\,III}]/([\mathrm{O\,II}]\times z)$, indicating luminous AGNs hosted by faint dwarfs.

Load-bearing premise

The interpretation of the two WISE clumps as AGN versus starburst contamination assumes the $[\mathrm{O\,II}]/z$ and $[\mathrm{O\,III}]/([\mathrm{O\,II}]\times z)$ proxy maps are faithful tracers of star formation and AGN strength, but the paper does not define the emission-line flux measurements, leaving open whether dust extinction or host dilution biases those proxies.

Editorial extensions

If this is right

  • BPT-selected dwarf AGN samples are systematically biased toward the high-mass end of the dwarf population, so occupation fractions built on BPT alone undercount low-mass, metal-poor hosts.
  • WISE-selected dwarf AGN samples contain a measurable starburst-contaminated subpopulation that can be identified by its position on the SED manifold, making the clump split a practical contamination filter.
  • Variability-selected AGNs distribute almost uniformly across the SED manifold, suggesting that upcoming time-domain surveys will deliver a much less host-biased dwarf AGN census than line- or infrared-selected samples.
  • Emission-line, broad-line, and WISE selection all avoid hosts with strong star formation, so the union of current diagnostics is missing a population of dwarf AGNs embedded in actively star-forming galaxies.
  • Cells with high $[\mathrm{O\,III}]/([\mathrm{O\,II}]\times z)$ that host AGNs in low-mass galaxies imply luminous AGNs hiding in faint dwarfs, recoverable by combining manifold position with line-ratio information.

Reading between the lines

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

  • The paper stops at mapping the clumps; a direct next step would be to convert the SOM into a per-cell completeness function, so that each diagnostic's recovery fraction can be integrated over the manifold to produce method-corrected AGN occupation fractions for dwarf galaxies.
  • The left/right WISE clump split implies that a simple cut in SOM coordinates—roughly a particular combination of near-UV to mid-infrared colors—could serve as an empirical contamination filter, although the paper does not prescribe the cut itself.
  • Because the SOM preserves topology, the same trained map could be applied to higher-redshift dwarf galaxies in deeper surveys, transferring the measured selection biases to regimes where current diagnostics are even less reliable.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 7 minor

Summary. The paper trains a 51×51 self-organizing map on 30,344 dwarf galaxies from the NSA catalog (z < 0.055, M* < 10^9.5 Msun) using seven colors from nine bands (GALEX FUV/NUV, SDSS ugriz, WISE W1/W2). A sample of 438 previously identified dwarf AGNs, compiled by Wasleske & Baldassare (2024), is mapped onto the SOM and the locations of AGNs selected by BPT, [O i], [S ii], He ii, broad-line, X-ray, variability, and WISE methods are compared. The central claim is that different AGN diagnostics occupy distinct and partially overlapping regions of SED space, reflecting host-property selection biases. The most novel finding is that WISE-selected AGNs form two clumps: a redder, more AGN-like left clump and a bluer, starburst-like right clump, suggesting a way to reduce starburst contamination in WISE-selected dwarf AGN samples.

Significance. If the central claims hold, the paper offers a useful data-driven visualization of AGN selection functions in the dwarf galaxy regime and a potential route to decontaminate WISE-selected dwarf AGN samples. The use of an external AGN compilation and the authors' explicit test of the WISE-input circularity are commendable. However, the most load-bearing interpretation—that the left WISE clump is AGN-dominated and the right clump is starburst-contaminated—rests on emission-line proxy maps that are not quantitatively defined, and the clump split itself is based on visual inspection without robustness tests. The paper therefore presents a plausible and interesting atlas, but the novel quantitative claims are not yet established to the standard required for a journal publication.

major comments (4)
  1. [§4.2, Figures 4 and 5] The central interpretation that the left WISE clump contains genuine AGNs while the right clump is contaminated by starbursts rests almost entirely on the [O ii]/z and [O iii]/([O ii]×z) maps. The manuscript never defines how the [O ii] and [O iii] fluxes are measured: are they from SDSS spectroscopy, are they continuum-subtracted, are dust-extinction corrections applied, and how are the line fluxes matched to the photometric apertures used for the SOM? Moreover, the [O iii]/([O ii]×z) ratio is structurally sensitive to low [O ii] flux: red, quiescent galaxies—exactly the population identified with the left clump—will show large values of this ratio even if their [O iii] flux is modest. The paper should report the underlying [O ii] and [O iii] fluxes or equivalent widths for the two clumps separately, and demonstrate the AGN nature of the left clump using a standard diagnostic such as [O iii]/Hβ or the BPT classification where available.
  2. [§4.2, Figure 6] The division of the 83 WISE-selected AGNs into 55 left-clump and 22 right-clump objects is based on visual inspection of the overplotted SOM without a quantitative definition of the clump boundaries. There is no specification of which cells belong to each clump, no test of sensitivity to SOM hyperparameters (grid size, initialization, training length), and no significance test that the two clumps are statistically distinct in their SED properties. A reproducible clump-membership criterion and a stability analysis of the split are needed before the two-component interpretation can be evaluated.
  3. [§4.2, page 11] The authors acknowledge that WISE bands are part of the SOM input and that this can enhance the apparent clustering of WISE-selected AGNs. They state that after retraining the SOM without WISE photometry, the WISE-selected sources still occupy broadly similar regions with a less concentrated clumpy structure, but they provide no figure, quantitative measure, or comparison of clump membership for this retrained map. Since the W1−W2 color is itself a WISE AGN selection criterion, the persistence of the large-scale structure is essential to separate a genuine host-property selection function from a built-in color cut. The retraining result should be shown and quantified.
  4. [§4.1 and §4.2] The comparison of spatial distributions across selection methods (e.g., the claims that X-ray and variability-selected AGNs are 'uniformly distributed' or that BPT-selected AGNs cluster in high-mass regions) is made by visual inspection, with no statistical test or uncertainty estimates. Given the small numbers in some categories, the authors should provide at least simple quantitative measures—such as the distribution of SOM cell densities relative to the training sample, or a two-dimensional Kolmogorov–Smirnov test—to support the stated differences between selection methods.
minor comments (7)
  1. [§2.1 vs §5] The NSA catalog version is given as nsa_v1_0_2 in §2.1 and as nsa_v1_0_1 in §5; please correct the inconsistency.
  2. [§5] The reference to Wasleske & Baldassare (2024) appears as 'compiled by ?' in the summary section; the citation needs to be completed.
  3. [§1 and title] There is a typo in 'low metallicites' in the introduction, and the title contains a spurious space in 'T echniques' due to formatting; please fix both.
  4. [Figure 2 caption] The caption begins with 'T op row' instead of 'Top row'; this typo should be corrected.
  5. [§2.2] The text says the cross-match used the WISE All-Sky Source catalog and the nine-band photometry includes only W1 and W2, but the earlier sentence mentions four WISE bands; please clarify why W3 and W4 are not used in the SOM input.
  6. [§4.3] The coronal-line emitter section gives 57 cross-matched sources and 48 unique ones, but the matching radius and selection details are not described; adding a sentence on the cross-match and completeness would help.
  7. [Figure 9] The color–mass diagrams would benefit from a brief description of how the training galaxies in the left/right clump cells are selected and how many galaxies each contour represents, since the current caption relies on visual comparison with Schawinski et al. (2014) reference lines.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity; the SOM analysis is unsupervised, AGN samples are external, and the acknowledged WISE-input dependence is checked by retraining without WISE.

full rationale

The derivation chain is self-contained rather than circular. The SOM is trained without labels on 30,344 NSA dwarf galaxies using seven photometric colors, and the 438 AGNs are mapped as an external test sample drawn from Wasleske & Baldassare (2024); no AGN label or selection criterion is used to train the map. The only place where an input feature coincides with a measured result is the concentration of WISE-selected AGNs in high W1-W2 SOM cells, and the paper explicitly acknowledges this: 'since W1−W2 was used as part of the input for training (Figure 1). As a result, the SOM mapping process is naturally sensitive to variations in this color.' Crucially, the authors do not stop there; they retrain the SOM without WISE photometry and find that the broad structure persists, so the central claim does not reduce to the input by construction. The two-clump interpretation is tested with independent information that was not part of the SOM training: the [O ii]/z and [O iii]/([O ii]×z) line-ratio maps, SDSS images, and u-r versus stellar mass diagrams. The line-ratio proxies are not fully defined in the text, which is a validity or support concern, but it is not circularity. There is no fitted parameter renamed as a prediction, no load-bearing self-citation chain, no imported uniqueness theorem, and no known result merely relabeled in new coordinates. The paper's claims are descriptive and externally anchored, so the circularity score is 0.

Assumptions & free parameters 3 free parameters · 5 assumptions · 0 invented entities

The central descriptive claims rest on a few modeling choices: the SOM grid and hyperparameters, the WISE cross-match tolerance, and the line-ratio proxy maps. No new physical entities are introduced. The grid size and clump boundaries are hand-chosen; the line-ratio maps are not formally specified.

free parameters (3)
  • SOM grid size = 51 x 51
    Chosen by elbow of quantization error curve; grid resolution controls apparent clumpiness and cell-level statistics.
  • SOM training hyperparameters = not stated (SOMPY defaults)
    Learning rate, neighborhood radius, and number of epochs are not reported despite affecting topology and cluster separation.
  • WISE clump boundaries = manual split: 55 left, 22 right
    Left and right clump assignment is by visual inspection; the headline counts and the starburst contamination claim depend on this hand-drawn division.
assumptions (5)
  • domain assumption The seven colors derived from nine bands span the SED manifold of dwarf galaxies.
    Used to train the SOM; if missing features, the manifold and cell neighborhoods would differ. Invoked in Section 3.2.
  • domain assumption WISE All-Sky cross-match at 2 arcsec, plus iterative matching up to 1e-2 degrees, associates the correct galaxies.
    Forms the 30,344 training sample and the 438 AGN matches. Sections 2.2 and 2.3.
  • domain assumption [O ii]/z and [O iii]/([O ii]*z) flux ratios are valid proxies for star formation and AGN luminosity without specified extinction corrections.
    This proxy underpins the starburst versus AGN interpretation of the WISE clumps. Figures 4 and 5, Section 4.1.
  • domain assumption The Wasleske and Baldassare (2024) compilation is a representative test set of dwarf AGNs for the diagnostics considered.
    All conclusions about selection biases are measured against this compilation. Section 2.3.
  • standard math The SOM, with 51x51 grid and SOMPY defaults, preserves the topology of the input color space.
    SOM theoretical guarantee; the paper checks only quantization error. Section 3.2.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Selection of Dwarf Galaxies Hosting AGNs: A Measure of Bias and Contamination using Unsupervised Machine Learning Techniques." pith.science (2026). https://pith.science/paper/YNKUYK7L

@misc{pith2026250516509,
  author       = {Pith},
  title        = {Pith review of: Selection of Dwarf Galaxies Hosting AGNs: A Measure of Bias and Contamination using Unsupervised Machine Learning Techniques},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/YNKUYK7L}},
  note         = {Machine review of arXiv:2505.16509}
}
read the original abstract

Identifying AGNs in dwarf galaxies is critical for understanding black hole formation but remains challenging due to their low luminosities, low metallicities, and star formation-driven emission that can obscure AGN signatures. Machine learning (ML) techniques, particularly unsupervised methods, offer new ways to address these challenges by uncovering patterns in complex data. In this study, we apply Self-Organizing Maps (SOMs) to explore the SED manifold of dwarf galaxies and evaluate AGN selection biases across diagnostics. We train a 51 by 51 SOM on 30,344 dwarf galaxies (redshift less than 0.055 and stellar mass below 10 to the 9.5 solar masses) from the NSA catalog using nine-band photometry from near-UV to mid-infrared. A set of 438 previously identified dwarf AGNs, selected via various methods, was mapped onto the SOM. AGNs identified by different methods occupy distinct and partially overlapping regions in SED space, reflecting selection biases tied to host properties. BPT selected AGNs cluster in higher-mass regions, while X-ray and variability-selected AGNs show broader distributions. WISE-selected AGNs are concentrated in lower-mass regions and form two clumps: one associated with bluer, starburst-like systems and the other with redder, more AGN-like SEDs. This separation may help distinguish true AGN hosts from starburst contaminants in WISE-selected samples. AGNs selected via traditional emission-line, broad-line, and WISE methods tend to avoid SOM regions linked to strong star formation. In contrast, a subset of AGNs in low-mass galaxies occupy regions indicative of high AGN luminosity relative to stellar content, highlighting luminous AGNs in faint hosts. These results demonstrate the utility of manifold learning for improving AGN selection in the low-mass regime.

Figures

Figures reproduced from arXiv: 2505.16509 by the authors.

Figure 1
Figure 1. Component maps of the trained SOM, showing the learned 7-dimensional color manifold. Each panel corresponds to one of the color dimensions, with the color bars indicating the variation in that color across the SOM cells. input features used during training, reflecting the struc￾ture learned from the photometric color space [PITH_FULL_IMAGE:figures/full_fig_p005_1.png] view at source ↗
Figure 2
Figure 2. SOM trained on the photometric colors of dwarf galaxies. Top row: SOM cells color-coded by the logarithmic median stellar mass (left) and photometric redshift (right) of the training galaxies mapped to each cell. Bottom row: SOM cells color-coded by the number density of training galaxies (left) and the median Hα/Hβ line ratio (right). White regions along the lower edges of the SOM correspond to cells where Hα or Hβ… view at source ↗
Figure 3
Figure 3. Distribution of NSA dwarf AGNs mapped onto the trained SOM, with cells color-coded by the logarithmic median stellar mass of the training dwarf galaxies. Pink squares indicate AGNs associated with representative selection methods. result, current AGN samples in dwarf galaxies are in￾complete and shaped by significant selection biases, lim￾iting efforts to robustly determine black hole occupation fractions or scaling… view at source ↗
Figures from the paper (7 more)
Figure 4
Figure 4. Figure 4: Distribution of NSA dwarf AGNs mapped onto the trained SOM, with cells color-coded by the [O ii]/z flux ratio of the training dwarf galaxies. The z corresponds to SDSS z-band photometry, tracing the stellar continuum near 9000 ˚A. Pink squares indicate AGNs associated …
Figure 5
Figure 5. Figure 5: Distribution of NSA dwarf AGNs mapped onto the trained SOM, with cells color-coded by the [O iii]/([O ii] × z) flux ratio of the training dwarf galaxies. The z corresponds to SDSS z-band photometry, tracing the stellar continuum near 9000 ˚A. Pink squares indicate AGNs…
Figure 6
Figure 6. Figure 6: WISE-selected dwarf AGNs mapped onto the trained SOM, shown as pink squares overlaid on cyan cir￾cles, which represent the full sample of NSA dwarf AGNs se￾lected by various methods. Two distinct WISE AGN clumps are visible: the right clump is associated with bluer sou…
Figure 7
Figure 7. Figure 7: Comparison of the stellar mass distributions of WISE-selected dwarf AGNs and those selected through other methods. WISE AGNs show a clear preference for lower￾mass hosts relative to the broader dwarf AGN sample. In general, as shown by the distribution of cyan circles …
Figure 8
Figure 8. Figure 8: SDSS images of a sample of WISE-selected AGNs from the left (top) and right (bottom) clumps. The scale is 0.4”/pixel. Galaxies in the left clump appear less blue, and several of them have a bright source in the center of the galaxy. Galaxies in the right clump appear b…
Figure 9
Figure 9. Figure 9: Color–mass diagrams of all training and test sample galaxies located within the left and right WISE clump cells of the SOM. The bluer u−r colors of the right clump galaxies suggest a population dominated by young starbursts, consistent with contamination in WISE-based …
Figure 10
Figure 10. Figure 10: Distribution of NSA dwarf AGNs and coronal line emitters mapped onto the trained SOM, which is color￾coded by the logarithmic median stellar mass of the train￾ing galaxies in each cell. Cyan circles indicate the mapped AGNs, while red crosses show the coronal line emi…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

62 extracted references · 7 canonical work pages

  1. [1]

    2024, The Astrophysical Journal, 975, 60, doi: 10.3847/1538-4357/ad702b

    Bohn, T. 2024, The Astrophysical Journal, 975, 60, doi: 10.3847/1538-4357/ad702b

    doi:10.3847/1538-4357/ad702b 2024

  2. [2]

    F., Geha, M., & Greene, J

    Baldassare, V. F., Geha, M., & Greene, J. 2018, The Astrophysical Journal, 868, 152, doi: 10.3847/1538-4357/aae6cf

    doi:10.3847/1538-4357/aae6cf 2018

  3. [3]

    M., & Terlevich, R

    Baldwin, A., Phillips, M. M., & Terlevich, R. 1981, Publications of the Astronomical Society of the Pacific, 93, 817, doi: 10.1086/130930

    doi:10.1086/130930 1981

  4. [4]

    G., Kocevski, D

    Barro, G., P´ erez-Gonz´ alez, P. G., Kocevski, D. D., et al. 2024, ApJ, 963, 128, doi: 10.3847/1538-4357/ad167e

    doi:10.3847/1538-4357/ad167e 2024

  5. [5]

    S., Wechsler, R

    Behroozi, P. S., Wechsler, R. H., & Conroy, C. 2013, The Astrophysical Journal, 770, 57, doi: 10.1088/0004-637X/770/1/57

    doi:10.1088/0004-637x/770/1/57 2013

  6. [6]

    L., Watson, M

    Birchall, K. L., Watson, M. G., Aird, J., & Starling, R. L. C. 2021, Monthly Notices of the Royal Astronomical Society, 510, 4556, doi: 10.1093/mnras/stab3573 AGN SOM 15

    doi:10.1093/mnras/stab3573 2021

  7. [7]

    2011, The Astronomical Journal, 142, 31, doi: 10.1088/0004-6256/142/1/31

    Price-Whelan, A. 2011, The Astronomical Journal, 142, 31, doi: 10.1088/0004-6256/142/1/31

    doi:10.1088/0004-6256/142/1/31 2011

  8. [8]

    2021, ApJ, 911, 70, doi: 10.3847/1538-4357/abe70c

    Bohn, T., Canalizo, G., Veilleux, S., & Liu, W. 2021, ApJ, 911, 70, doi: 10.3847/1538-4357/abe70c

    doi:10.3847/1538-4357/abe70c 2021

Show all 62 references
  1. [9]

    J., Shen, Y., Blaes, O., et al

    Burke, C. J., Shen, Y., Blaes, O., et al. 2021, Science, 373, 789, doi: 10.1126/science.abg9933

    2021 doi

  2. [10]

    M., Satyapal, S., Abel, N

    Cann, J. M., Satyapal, S., Abel, N. P., et al. 2019, The Astrophysical Journal Letters, 870, L2, doi: 10.3847/2041-8213/aaf88d

    2019 doi

  3. [11]

    R., et al

    Chartab, N., Mobasher, B., Cooray, A. R., et al. 2023, The Astrophysical Journal, 942, 91, doi: 10.3847/1538-4357/acacf5

    2023 doi

  4. [12]

    L., et al

    Davidzon, I., Laigle, C., Capak, P. L., et al. 2019, MNRAS, 489, 4817, doi: 10.1093/mnras/stz2486

    2019 doi

  5. [13]

    2022, A&A, 665, A34, doi: 10.1051/0004-6361/202243249

    Davidzon, I., Jegatheesan, K., Ilbert, O., et al. 2022, A&A, 665, A34, doi: 10.1051/0004-6361/202243249

    2022 doi

  6. [14]

    2016, Monthly Notices of the Royal Astronomical Society, 456, 3354, doi: 10.1093/mnras/stv2794

    Feltre, A., Charlot, S., & Gutkin, J. 2016, Monthly Notices of the Royal Astronomical Society, 456, 3354, doi: 10.1093/mnras/stv2794

    2016 doi

  7. [15]

    Geach, J. E. 2012, Monthly Notices of the Royal Astronomical Society, 419, 2633, doi: 10.1111/j.1365-2966.2011.19913.x

    2012

  8. [16]

    E., Strader, J., & Ho, L

    Greene, J. E., Strader, J., & Ho, L. C. 2020, ARA&A, 58, 257, doi: 10.1146/annurev-astro-032620-021835

    2020 doi

  9. [17]

    E., Labbe, I., Goulding, A

    Greene, J. E., Labbe, I., Goulding, A. D., et al. 2024, ApJ, 964, 39, doi: 10.3847/1538-4357/ad1e5f

    2024 doi

  10. [18]

    A., Kocevski, D

    Grogin, N. A., Kocevski, D. D., Faber, S., et al. 2011, The Astrophysical Journal Supplement Series, 197, 35

    2011

  11. [20]

    E., Siegmund, W

    Gunn, J. E., Siegmund, W. A., Mannery, E. J., et al. 2006, The Astronomical Journal, 131, 2332, doi: 10.1086/500975

    2006 doi

  12. [21]

    N., Reines, A

    Hainline, K. N., Reines, A. E., Greene, J. E., & Stern, D. 2016, The Astrophysical Journal, 832, 119, doi: 10.3847/0004-637X/832/2/119

    2016 doi

  13. [22]

    2023, ApJ, 959, 39, doi: 10.3847/1538-4357/ad029e

    Harikane, Y., Zhang, Y., Nakajima, K., et al. 2023, ApJ, 959, 39, doi: 10.3847/1538-4357/ad029e

    2023 doi

  14. [23]

    2005, The Astrophysical Journal, 634, 161–168, doi: 10.1086/497047

    Kauffmann, G. 2005, The Astrophysical Journal, 634, 161–168, doi: 10.1086/497047

    2005 doi

  15. [24]

    2019, ApJL, 881, L14, doi: 10.3847/2041-8213/ab3418

    Hemmati, S., Capak, P., Pourrahmani, M., et al. 2019, ApJL, 881, L14, doi: 10.3847/2041-8213/ab3418

    2019 doi

  16. [25]

    2019, The Astrophysical Journal, 877, 117, doi: 10.3847/1538-4357/ab1be5

    Hemmati, S., Capak, P., Masters, D., et al. 2019, The Astrophysical Journal, 877, 117, doi: 10.3847/1538-4357/ab1be5

    2019 doi

  17. [26]

    C., Jones, C., Forman, W

    Hickox, R. C., Jones, C., Forman, W. R., et al. 2009, The Astrophysical Journal, 696, 891, doi: 10.1088/0004-637X/696/1/891

    2009 doi

  18. [27]

    2017, Monthly Notices of the Royal Astronomical Society, 472, 2468, doi: 10.1093/mnras/stx2130

    Hirschmann, M., Groves, B., Schaye, J., et al. 2017, Monthly Notices of the Royal Astronomical Society, 472, 2468, doi: 10.1093/mnras/stx2130

    2017 doi

  19. [28]

    H., Cohen, M., Masci, F., et al

    Jarrett, T. H., Cohen, M., Masci, F., et al. 2011, The Astrophysical Journal, 735, 112, doi: 10.1088/0004-637X/735/2/112

    2011 doi

  20. [29]

    H., Masci, F., Tsai, C

    Jarrett, T. H., Masci, F., Tsai, C. W., et al. 2013, AJ, 145, 6, doi: 10.1088/0004-6256/145/1/6

    2013 doi

  21. [30]

    M., Tremonti, C., et al

    Kauffmann, G., Heckman, T. M., Tremonti, C., et al. 2003, Monthly Notices of the Royal Astronomical Society, 346, 1055–1077, doi: 10.1111/j.1365-2966.2003.07154.x

    2003

  22. [31]

    1998, ARA&A, 36, 189, doi: 10.1146/annurev.astro.36.1.189

    Kennicutt, Robert C., J. 1998, ARA&A, 36, 189, doi: 10.1146/annurev.astro.36.1.189

    1998 doi

  23. [32]

    2001, The Astrophysical Journal, 556, doi: 10.1086/321545

    Trevena, J. 2001, The Astrophysical Journal, 556, doi: 10.1086/321545

    2001 doi

  24. [33]

    J., Groves, B., Kauffmann, G., & Heckman, T

    Kewley, L. J., Groves, B., Kauffmann, G., & Heckman, T. 2006, MNRAS, 372, 961, doi: 10.1111/j.1365-2966.2006.10859.x

    2006

  25. [34]

    D., Onoue, M., Inayoshi, K., et al

    Kocevski, D. D., Onoue, M., Inayoshi, K., et al. 2023, ApJL, 954, L4, doi: 10.3847/2041-8213/ace5a0

    2023 doi

  26. [35]

    D., Finkelstein, S

    Kocevski, D. D., Finkelstein, S. L., Barro, G., et al. 2024, arXiv e-prints, arXiv:2404.03576, doi: 10.48550/arXiv.2404.03576

    2024 doi arXiv

  27. [36]

    M., Faber, S

    Koekemoer, A. M., Faber, S. M., Ferguson, H. C., et al. 2011, ApJS, 197, 36, doi: 10.1088/0067-0049/197/2/36

    2011 doi

  28. [37]

    1982, Biological Cybernetics, doi: 10.1007/BF00337288

    Kohonen, T. 1982, Biological Cybernetics, doi: 10.1007/BF00337288

    1982 doi

  29. [38]

    E., Reddy, N

    Kriek, M., Shapley, A. E., Reddy, N. A., et al. 2015, ApJS, 218, 15, doi: 10.1088/0067-0049/218/2/15 La Torre, V., Sajina, A., Goulding, A. D., et al. 2024, AJ, 167, 261, doi: 10.3847/1538-3881/ad3821 Labb´ e, I., van Dokkum, P., Nelson, E., et al. 2023, Nature, 616, 266, doi:...

    2015 doi

  30. [39]

    E., Bezanson, R., et al

    Labbe, I., Greene, J. E., Bezanson, R., et al. 2025, ApJ, 978, 92, doi: 10.3847/1538-4357/ad3551

    2025 doi

  31. [40]

    2017, MNRAS, 467, 540, doi: 10.1093/mnras/stx055

    Lamperti, I., Koss, M., Trakhtenbrot, B., et al. 2017, MNRAS, 467, 540, doi: 10.1093/mnras/stx055

    2017 doi

  32. [41]

    2024, A&A, 691, A145, doi: 10.1051/0004-6361/202347640

    Maiolino, R., Scholtz, J., Curtis-Lake, E., et al. 2024, A&A, 691, A145, doi: 10.1051/0004-6361/202347640

    2024 doi

  33. [42]

    2015, The Astrophysical Journal, 813, 53, doi: 10.1088/0004-637X/813/1/53

    Masters, D., Capak, P., Stern, D., et al. 2015, The Astrophysical Journal, 813, 53, doi: 10.1088/0004-637X/813/1/53

    2015 doi

  34. [43]

    P., Brammer, G., et al

    Matthee, J., Naidu, R. P., Brammer, G., et al. 2024, ApJ, 963, 129, doi: 10.3847/1538-4357/ad2345 16 Sanjaripour et al

    2024 doi

  35. [44]

    2013, ApJ, 770, 40, doi: 10.1088/0004-637X/770/1/40

    Hopkins, A. 2013, ApJ, 770, 40, doi: 10.1088/0004-637X/770/1/40

    2013 doi

  36. [45]

    2021, ApJ, 922, 155, doi: 10.3847/1538-4357/ac1ffa

    Salehirad, S. 2021, ApJ, 922, 155, doi: 10.3847/1538-4357/ac1ffa

    2021 doi

  37. [46]

    2014, SOMPY: A Python Library for Self Organizing Map (SOM) M¨ uller-S´ anchez, F., Prieto, M

    Moosavi, V., Packmann, S., & Vall´ es, I. 2014, SOMPY: A Python Library for Self Organizing Map (SOM) M¨ uller-S´ anchez, F., Prieto, M. A., Hicks, E. K. S., et al. 2011, ApJ, 739, 69, doi: 10.1088/0004-637X/739/2/69

    2014 doi

  38. [47]

    B., & Gunn, J

    Oke, J. B., & Gunn, J. E. 1983, ApJ, 266, 713, doi: 10.1086/160817

    1983 doi

  39. [48]

    O., et al

    Reefe, M., Satyapal, S., Sexton, R. O., et al. 2022, ApJ, 936, 140, doi: 10.3847/1538-4357/ac8981

    2022 doi

  40. [49]

    O., Doan, S

    Reefe, M., Sexton, R. O., Doan, S. M., et al. 2023, The Astrophysical Journal Supplement Series, 265, 21, doi: 10.3847/1538-4365/acb0d2

    2023 doi

  41. [50]

    E., Greene, J

    Reines, A. E., Greene, J. E., & Geha, M. 2013, The Astrophysical Journal, 775, 116, doi: 10.1088/0004-637X/775/2/116

    2013 doi

  42. [51]

    E., & Volonteri, M

    Reines, A. E., & Volonteri, M. 2015, ApJ, 813, 82

    2015

  43. [52]

    Riffel, R., Rodr ´ ıguez-Ardila, A., & Pastoriza, M. G. 2006, A&A, 457, 61, doi: 10.1051/0004-6361:20065291 Rodr ´ ıguez-Ardila, A., Prieto, M. A., Portilla, J. G., &

    2006 doi

  44. [54]

    2002, ApJ, 579, 214, doi: 10.1086/342840 Rodr ´ ıguez-Ardila, A., Pastoriza, M

    Prato, L. 2002, ApJ, 579, 214, doi: 10.1086/342840 Rodr ´ ıguez-Ardila, A., Pastoriza, M. G., Viegas, S., Sigut, T. A. A., & Pradhan, A. K. 2002, The Astrophysical Journal, 579, 214, doi: 10.1086/342747 Rodr ´ ıguez-Ardila, A., Prieto, M. A., Portilla, J. G., &

    2002 doi

  45. [55]

    Tejeiro, J. M. 2011, The Astrophysical Journal, 743, 100, doi: 10.1088/0004-637X/743/2/100

    2011 doi

  46. [56]

    2024, The Astrophysical Journal, 977, 202, doi: 10.3847/1538-4357/ad90ba

    Sanjaripour, S., Hemmati, S., Mobasher, B., et al. 2024, The Astrophysical Journal, 977, 202, doi: 10.3847/1538-4357/ad90ba

    2024 doi

  47. [57]

    J., McAlpine, W., et al

    Satyapal, S., Secrest, N. J., McAlpine, W., et al. 2014, The Astrophysical Journal, 784, 113, doi: 10.1088/0004-637X/784/2/113

    2014 doi

  48. [58]

    M., Simmons, B

    Schawinski, K., Urry, C. M., Simmons, B. D., et al. 2014, MNRAS, 440, 889, doi: 10.1093/mnras/stu327

    2014 doi

  49. [59]

    J., Benford, D

    Stern, D., Assef, R. J., Benford, D. J., et al. 2012, The Astrophysical Journal, 753, 30, doi: 10.1088/0004-637X/753/1/30 S´ anchez-S´ aez, P., Lira, P., Cartier, R., & et al. 2018, Astronomy & Astrophysics, 614, A40, doi: 10.1051/0004-6361/201732228

    2012 doi

  50. [60]

    R., et al

    Trump, J. R., et al. 2011, The Astrophysical Journal, 733, 60, doi: 10.1088/0004-637X/733/1/60

    2011 doi

  51. [61]

    R., Sun, M., Zeimann, G

    Trump, J. R., Sun, M., Zeimann, G. R., et al. 2015, The Astrophysical Journal, 811, 26, doi: 10.1088/0004-637X/811/1/26 van den Busch, J. L., Wright, A. H., Hildebrandt, H., et al. 2022, A&A, 664, A170, doi: 10.1051/0004-6361/202142083 van der Marel, R. P., Alves, D. R., Hardy...

    2015 doi

  52. [62]

    J., & Baldassare, V

    Wasleske, E. J., & Baldassare, V. F. 2024, The Astrophysical Journal, 921, 68, doi: 10.3847/1538-4357/ad5442

    2024 doi

  53. [63]

    L., Eisenhardt, P

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

    2010 doi

  54. [64]

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

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

    2000 doi

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.