Combining summary statistics with simulation-based inference for the 21 cm signal from the Epoch of Reionization

Ashutosh Mishra; Benoit Semelin; David Cornu; Romain M\'eriot

REVIEW 4 major objections 6 minor 44 references

Combining summary statistics with simulation-based inference for the 21 cm signal from the Epoch of Reionization

T0 review · 4 major / 6 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read The paper claims that simulation-based inference can combine the 21 cm power spectrum with linear moments of the pixel distribution function to tighten constraints on reionization parameters in most cases, without an analytic likelihood.

desk verdict Solid SBI methodology for combining 21 cm summary statistics, but the headline 4D-contraction claim needs a higher-dimensional coverage test before it can be trusted. read the letter →

arxiv 2411.14419 v1 pith:BVYPEP3J submitted 2024-11-21 astro-ph.CO

Benoit Semelin , Romain Mériot , Ashutosh Mishra , David Cornu This is my paper · ORCID

classification astro-ph.CO

keywords 21cmcosmologyEpochofReionizationsimulation-basedinferenceneuraldensityestimatorspowerspectrumpixeldistributionfunctionL-momentsBayesian

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

With the upcoming low-frequency radio telescope expected to produce tomographic maps of the 21 cm signal from the Epoch of Reionization, the practical question is how to squeeze the most astrophysical information out of those maps. This paper argues that one route is to combine two imperfect summary statistics — the power spectrum and linear moments of the pixel distribution function — and let simulation-based inference learn their joint behavior directly from radiative-hydrodynamics simulations, no analytic likelihood required. Running roughly 900 inferences across the parameter space, the authors find the combined statistic shrinks the 4-D posterior volume in 91.5% of cases and tightens the individual 1-D parameter posteriors in 70–80% of cases, with median contractions of a factor of a few and 20–30%, respectively. They also report that the learned posteriors are biased by no more than about 20% of their standard deviation and under-confident by no more than about 15%. The authors frame the result as a practical alternative to the search for theoretically sufficient statistics.

What carries the argument

The load-bearing object is a neural density estimator that outputs the parameters of a two-component Gaussian mixture model for the likelihood of the summary statistics given the model parameters. The network is trained on a large set of parameter–summary pairs produced by running radiative-hydrodynamics simulations with many independent realizations of instrumental noise; training minimizes the negative log-likelihood of the samples, equivalently the Kullback–Leibler divergence between the true and learned likelihoods. Each Gaussian component's covariance matrix is parameterized through the Cholesky factor of its inverse, guaranteeing positive definiteness, and the same architecture ingests the power-spectrum vector, the linear-moment vector, or their concatenation. This learned joint likelihood is then used inside a Markov-chain Monte Carlo sampler, so the combination of statistics automatically accounts for the correlations between the power spectrum and the pixel moments.

What would settle it

Repeat the same 908-inference comparison with a more flexible representation of the likelihood, such as a normalizing flow: if the 91.5% contraction rate and the median factor-of-a-few volume shrinkage do not persist, the headline gain is an artifact of the Gaussian mixture approximation. A coverage test in the full 4-D parameter space would also settle whether the volume claim is trustworthy.

Watch

Extended reading notes

Core claim

The central result is that combining the isotropic 21 cm power spectrum with the linear moments $l_2$–$l_6$ of the pixel distribution function — two statistics that are correlated and have no joint analytic likelihood — tightens Bayesian constraints on reionization parameters compared with using either alone. After training neural density estimators to model a two-component Gaussian mixture likelihood and running 908 inferences at different locations in parameter space, the paper reports that the combined summary contracts the 4-D posterior volume, computed from the square root of the generalized variance, in 91.5% of cases, and contracts individual marginalized 1-D posteriors in 70–80% of cases depending on the parameter. The median contraction is a factor of a few in the 4-D volume and 20–30% in the 1-D standard deviations. Simulation-based calibration indicates the 1-D marginalized posteriors are biased by no more than about 20% of their standard deviation and under-confident by no more than about 15%, except for the least constrained parameter ($f_{\rm esc}$), which is poorly sampled in the training database and whose posterior is markedly under-confident.

Load-bearing premise

The load-bearing premise is that the neural network's simplified two-component Gaussian description of how the statistics scatter around their true values is accurate enough that the reported shrinkage of the 4-D error volume reflects real information gain, even though the validation only looked at one parameter at a time.

Editorial extensions

If this is right

For a fixed observation setup, combining the power spectrum with linear PDF moments will typically yield tighter parameter constraints than the power spectrum alone, with the largest gain appearing in the 4-D posterior volume.
Because the neural density estimator learns correlations directly, the same combination strategy can be applied to other correlated summary statistics without assuming independence.
The linear moments $l_2$–$l_6$ outperform ordinary statistical moments $m_2$–$m_6$ both in calibration and in constraining power, making L-moments the better default for this kind of one-point statistic.
The occasional losses from combination, about 8.5% of cases for the 4-D volume and mostly when the power spectrum alone wins, are the exception rather than the rule, so combined inference can serve as a safe general-purpose strategy.

Reading between the lines

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

One implication the authors leave implicit: if the median 1-D tightening of 20–30% holds with real data, combining statistics is comparable to a moderate extension of observing time, but at zero additional telescope cost — a trade-off worth testing with a forecasting study.
Because simulation-based calibration only checks 1-D marginals, the headline 4-D volume contraction is not directly certified; a coverage test in the full 4-D space would be a sharper test of whether the volume gain is real.
The weak constraint on $f_{\rm esc}$ is tied to having only three sampled values in the database; adding more $f_{\rm esc}$ values or a lower-redshift snapshot could turn the combined statistic into a useful probe of photon escape, which the paper notes may help.

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.

Desk Editor's Note

read the letter

You asked for my take on the Semelin et al. SBI paper. The stress-test note is right, and the reader's conditional verdict is about where I land too.

What's genuinely new: this is the first quantified demonstration that SBI can combine the 21 cm power spectrum and linear PDF moments on a full radiative-hydrodynamics database, with ~900 inferences and an SBC-based assessment. That's a substantial computational effort. They compare four NDEs, report honestly that fesc is under-confident (fixed in the test sample) and that ordinary moments underperform L-moments, and they discuss model limitations in the conclusion. The use of L-moments is a sensible choice to avoid outlier sensitivity. The database is public, and the pipeline is reproducible in principle.

The soft spot is the validation gap, and it's real. The SBC is only on 1D marginalized posteriors; the authors themselves point out that two different joint posteriors can share the same 1D marginals. The headline 91.5% contraction is a 4D statement, measured on posteriors never validated above 1D. On top of that, each parameter point in Loreli II has a single cosmic-variance realization, so the NDE cannot directly learn the joint scatter that cosmic variance induces in the PS and PDF moments. If the 2-component GMM underestimates that correlated scatter, the 4D covariance determinant will be artificially small. A 4D coverage test—e.g. rank statistics on the posterior covariance or a multivariate SBC—would settle this, and I don't see a reason it can't be done.

One thing neither the reader nor the stress-test flagged: it's not specified whether the 908 SBC simulations are held out from the 90% training set. If they overlap, the calibration is optimistic. The authors should state this explicitly.

None of this sinks the paper. The approach is plausible, the code seems sound, and the finding that combining statistics tightens constraints is likely true in many cases. But the quantitative headline needs qualification and a higher-dimensional calibration before it becomes a benchmark. I'd send this to a serious referee, asking for those additions. I'd cite it in a methods context with the caveat. It's a good reading-group paper on the pitfalls of SBI validation.

Referee Report

4 major / 6 minor

Summary. The paper presents a simulation-based inference (SBI) pipeline for 21 cm Epoch of Reionization parameters, trained on the Loreli II database with LICORICE simulations. Three neural density estimators (NDEs) model the likelihood of the power spectrum, the linear moments of the pixel distribution function, and their combination, using a two-component Gaussian mixture model. The authors perform 908 inferences at interior parameter points, validate the 1D marginalized posteriors with simulation-based calibration (SBC), and report that the posteriors are biased by no more than about 20% of their standard deviation and under-confident by no more than about 15%. They then compare posterior volumes and find that combining the power spectrum with the linear PDF moments contracts the 4D generalized variance in 91.5% of cases and contracts the 1D marginalized widths in 70-80% of cases, with median contractions of a factor of a few in 4D and 20-30% in 1D. The central claim is that SBI can effectively combine non-Gaussian summary statistics to tighten EoR parameter constraints relative to either statistic alone.

Significance. If the reported volume contraction is genuine, this is a valuable and timely contribution for SKA-era 21 cm inference: it offers a practical route to combine non-Gaussian summary statistics without constructing sufficient statistics analytically. The study is computationally substantial, using about 9800 simulations, roughly 8 million noise-realised training samples, and 908 inferences, and the summary statistics are publicly available. The choice of L-moments is well motivated, and the paper explicitly tests calibration rather than assuming the learned likelihood is correct. I found no indication that the headline contraction is circular: the NDEs are trained on forward simulations, and the contraction is measured from the resulting approximate posteriors. The main weakness is that the headline 4D claim is validated only through 1D marginal tests, which are insensitive to exactly the kind of direction-selective approximation error that could affect the reported 4D volume contraction. This is addressable with additional validation, which is why I recommend major revision rather than rejection.

major comments (4)

[§3.3, §4.2] The central claim of a 91.5% contraction of the 4D posterior volume is not backed by a calibration test in the same dimension. The SBC validation in §3.3 is performed on 1D marginalized posteriors, and the text itself notes that different higher-dimensional posteriors can share the same 1D marginals. A 1D-calibrated two-component GMM may still be overconfident in particular directions of the 4D parameter space, and the generalized-variance ratio in Fig. 5 is exactly sensitive to such directions. I ask for a multivariate coverage diagnostic—for example, rank-based SBC applied to a 4D summary such as copula depth, or posterior-predictive checks of the joint covariance—before the headline contraction is claimed.
[§2.2, §3.2.2] The training data contain one cosmic-variance realization per parameter point and 1000 thermal-noise realizations per simulation. The NDE therefore learns the thermal-noise contribution to the likelihood covariance but not the cosmic-variance contribution to the joint scatter of the power-spectrum and PDF-moment summaries. For the combined-statistics posterior, which is built from the joint covariance between these summaries, an under-estimated cross-covariance will directly shrink det(Σ) and can produce a spurious apparent information gain. The 1D SBC histograms are insensitive to this kind of direction-selective overconfidence. At minimum, the paper should add a multivariate coverage test; ideally, it should also include a small set of independent cosmic-variance realizations at a few parameter points to check the learned joint covariance directly.
[§3.3, §4] The quantitative claims are conditional on a restricted test set. The 908 SBC simulations exclude the prior boundaries, and because only three fesc values exist in the database, this exclusion fixes fesc to a single value. Thus the statements that the posteriors are biased by at most 20% and under-confident by at most 15%, together with the 91.5% and 70-80% contraction percentages, have been demonstrated only for the interior of the prior with fesc fixed. The abstract and conclusions should state this restriction explicitly, or the test set should be extended by including boundary simulations or additional fesc sampling.
[§3.3] The 20% and 15% accuracy values are read off by eye from the SBC histograms by comparing them with toy normal models, and the text in §3.3 and §5 says '20% of the variance' where the toy model is a shift of 0.2 in a unit-variance normal, i.e., 20% of the standard deviation. Please replace the visual comparison with a quantitative SBC statistic (for example, the maximum absolute deviation from flatness of the rank histograms, or the Talts et al. rank statistic) and correct the variance/standard-deviation wording throughout.

minor comments (6)

[§5 vs. Abstract and Fig. 5] The conclusion says 'in 90% of the cases' while the abstract and Fig. 5 imply 91.5%; please reconcile the numbers.
[§3.3] Please state explicitly that the 908 SBC simulations are disjoint from the NDE training set; the 90/10 split makes this probable, but the text does not say so.
[§2.2, §3.3] There are several typos, including 'in particuliar' in §2.2 and 'maginalizations' in §3.3; a careful proofread is needed.
[Table 1] The caption says 'The average are computed'; moreover, the median values alone hide the large case-to-case variation visible in Figs. 5-6, so reporting the 16th-84th percentile spread of V and σ would be more informative.
[Figs. 5 and 6] The captions should define the regions where the combination outperforms the individual statistics; the text refers to a 'green region' in Fig. 5 that is not described in the caption.
[§4.2] Because some posteriors are bimodal or skewed (see Fig. 4d), reporting the standard deviation as the 1D measure of constraint can be misleading; consider adding the 68% interval width as a cross-check.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the posterior-contraction claim is an empirical measurement from forward simulations, not a reduction to fitted inputs or a self-citation chain.

full rationale

The paper's central claim—that combining the power spectrum and PDF linear moments contracts the 4-D posterior volume in 91.5% of 908 inferences—is obtained by training three NDEs on the same Loreli II forward simulations and comparing MCMC-derived posterior covariance matrices. No equation defines the claimed gain in terms of a fitted parameter; the 8.5% of cases where combining statistics inflates the volume shows the result is not forced by construction. The choice of a 2-component GMM is justified by prior work (Meriot et al. 2024; Prelogović & Mesinger 2023) and by the authors' own convergence tests, but this choice does not by itself guarantee posterior contraction. The SBC validation covers only 1D marginals, so the 4D contraction could in principle be an NDE approximation artifact; that is a calibration and correctness risk, not circularity. The paper itself flags that different joint posteriors can share identical 1D marginalizations, acknowledging the limitation rather than importing a uniqueness theorem. Self-citations concern the LICORICE/SPINTER code, the Loreli II database, and architecture comparisons; these support the inputs of the study, not the output claim. The derivation chain from forward simulations to posterior volumes is therefore self-contained with respect to the specific claim being made, and no circular step can be exhibited with the required reduction.

Assumptions & free parameters 5 free parameters · 6 assumptions · 0 invented entities

The central claim rests on choices of summary statistics, the Gaussian mixture likelihood parameterization, the Loreli II training set, and the noise model. These are standard SBI ingredients rather than derived facts; no new physical entities are introduced.

free parameters (5)

Power spectrum binning and k-range = 5 bins over k = 0.03 to 0.5 h/Mpc; 2 lowest and 1 highest bins discarded
Chosen to balance cosmic variance and thermal noise; the reported information gain may depend on this hand-selected compression.
PDF linear moments = l2 to l6 (5 moments)
Chosen to match the number of power spectrum bins; not optimized, and alternative moments m2 to m6 were found to perform worse.
Redshifts used in joint likelihood = z = 8.18, 10.32, 12.06
Selected to span the signal evolution in Loreli II; the authors note adding lower redshifts may improve fesc constraints.
Gaussian mixture components = Nc = 2
Chosen because no improvement was seen beyond two components; this limits the expressivity of the learned likelihood.
NDE training hyperparameters = 64 neurons per layer, tanh activation, batch size 4, initial learning rate 2e-4, two-stage training
Selected empirically by validation loss; these choices affect the accuracy of the approximate posterior and thus the measured gains.

assumptions (6)

domain assumption The true likelihood of the summary statistics is well approximated by a two-component Gaussian Mixture Model with parameter-dependent means, weights, and covariances.
Introduced in Section 3.2.2; validated only through SBC of 1D marginal posteriors, not the joint distribution.
domain assumption Each parameter point is represented by a single cosmic-variance realization in Loreli II, so cosmic variance is not averaged over in training.
Noted in Section 2.2; the 1000 thermal noise realizations per simulation cannot fully sample cosmic variance.
domain assumption The flat prior over the sampled Loreli II volume, which is not a hypercube and has correlated (tau, Mmin) pairs, is appropriate.
Assumed in Section 3.2.1; this prior determines the posterior and the SBC interpretation.
domain assumption The McQuinn et al. (2006) thermal noise model with the 2016 SKA baseline design is a faithful representation of future observations.
Used in Section 2.2 to generate noised cubes; the authors note the current baseline differs slightly.
domain assumption Downsampling the 21 cm cubes to 32^3 before computing summary statistics preserves the information relevant for parameter inference.
Motivated by the 6 arcmin noise resolution in Section 2.3; this removes small-scale signal information.
standard math The NDE trained by minimizing the negative log-likelihood converges to the true conditional likelihood in the limit of sufficient training data and expressivity.
Standard SBI principle stated in Section 3.1; no formal convergence guarantee is provided.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Combining summary statistics with simulation-based inference for the 21 cm signal from the Epoch of Reionization." pith.science (2026). https://pith.science/paper/BVYPEP3J

@misc{pith2026241114419,
  author       = {Pith},
  title        = {Pith review of: Combining summary statistics with simulation-based inference for the 21 cm signal from the Epoch of Reionization},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/BVYPEP3J}},
  note         = {Machine review of arXiv:2411.14419}
}

abstract

The 21 cm signal from the Epoch of Reionization will be observed with the up-coming Square Kilometer Array (SKA). SKA should yield a full tomography of the signal which opens the possibility to explore its non-Gaussian properties. How can we extract the maximum information from the tomography and derive the tightest constraint on the signal? In this work, instead of looking for the most informative summary statistics, we investigate how to combine the information from two sets of summary statistics using simulation-based inference. To this purpose, we train Neural Density Estimators (NDE) to fit the implicit likelihood of our model, the LICORICE code, using the Loreli II database. We train three different NDEs: one to perform Bayesian inference on the power spectrum, one to do it on the linear moments of the Pixel Distribution Function (PDF) and one to work with the combination of the two. We perform $\sim 900$ inferences at different points in our parameter space and use them to assess both the validity of our posteriors with Simulation-based Calibration (SBC) and the typical gain obtained by combining summary statistics. We find that our posteriors are biased by no more than $\sim 20 \%$ of their standard deviation and under-confident by no more than $\sim 15 \%$. Then, we establish that combining summary statistics produces a contraction of the 4-D volume of the posterior (derived from the generalized variance) in 91.5 % of our cases, and in 70 to 80 % of the cases for the marginalized 1-D posteriors. The median volume variation is a contraction of a factor of a few for the 4D posteriors and a contraction of 20 to 30 % in the case of the marginalized 1D posteriors. This shows that our approach is a possible alternative to looking for sufficient statistics in the theoretical sense.

Figures

Figures reproduced from arXiv: 2411.14419 by the authors.

**Figure 1.** Pixel distribution functions of all the simulations in the Loreli II database, illustrating the variety of possible shapes. The [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗

**Figure 2.** Architecture of the neural network used for the NDEs. The inputs are the model parameters and the summary statistics of a [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗

**Figure 3.** Histograms over the 908 inferences describing where the true parameter values fall in the cumulative distribution function of [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗

Figures from the paper (3 more)

**Figure 4.** Figure 4: Corner plot of the inferences done with the 3 NDEs on the di [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]

**Figure 5.** Figure 5: Scatter plot of the ratios of the characteristic volume oc [PITH_FULL_IMAGE:figures/full_fig_p009_5.png]

**Figure 6.** Figure 6: Same as fig. 5 but for the ratios of the volumes (i.e. standard deviation here) occupied by the 1D marginalized posteriors. [PITH_FULL_IMAGE:figures/full_fig_p010_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

44 extracted references · 32 canonical work pages

[1]

E., Alexander, P., et al

Abdurashidova, Z., Aguirre, J. E., Alexander, P., et al. 2022, ApJ, 924, 51

2022
[2]

2024, MNRAS, 534, L30

Acharya, A., Mertens, F., Ciardi, B., et al. 2024, MNRAS, 534, L30

work page 2024
[3]

2019, MNRAS, 488, 4440

Alsing, J., Charnock, T., Feeney, S., & Wandelt, B. 2019, MNRAS, 488, 4440

2019
[4]

& Wandelt, B

Alsing, J. & Wandelt, B. 2018, MNRAS, 476, L60

work page 2018
[5]

2009, A&A, 495, 389

Baek, S., Di Matteo, P., Semelin, B., Combes, F., & Revaz, Y . 2009, A&A, 495, 389

work page 2009
[6]

2010, A&A, 523, A4+

Baek, S., Semelin, B., Di Matteo, P., Revaz, Y ., & Combes, F. 2010, A&A, 523, A4+

work page 2010
[7]

& Hut, P

Barnes, J. & Hut, P. 1986, Nature, 324

work page 1986
[8]

D., Rogers, A

Bowman, J. D., Rogers, A. E. E., Monsalve, R. A., Mozdzen, T. J., & Mahesh, N. 2018, Nature, 555, 67

2018

Show all 44 references

[9]

Charnock, T., Lavaux, G., & Wandelt, B. D. 2018, Phys. Rev. D, 97, 083004

2018
[10]

& Zheng, Z

Chuzhoy, L. & Zheng, Z. 2007, ApJ, 670, 912

2007
[11]

Ciardi, B., Ferrara, A., & White, S. D. M. 2003, MNRAS, 344, 7

2003
[12]

& Weare, J

Goodman, J. & Weare, J. 2010, Communications in Applied Mathematics and Computational Science, 5, 65

2010
[13]

& Pritchard, J

Gorce, A. & Pritchard, J. R. 2019, MNRAS, 489, 1321

2019
[14]

& Mesinger, A

Greig, B. & Mesinger, A. 2015, MNRAS, 449, 4246

2015
[15]

Greig, B., Ting, Y .-S., & Kaurov, A. A. 2022, MNRAS, 513, 1719

2022
[16]

2009, MNRAS, 397, 1138

Harker, G., Zaroubi, S., Bernardi, G., et al. 2009, MNRAS, 397, 1138

2009
[17]

Hosking, J. R. M. 1990, Journal of the Royal Statistical Society, B52, 105

1990
[18]

Hosking, J. R. M. 1996, RC20525, B52, 105

1996
[19]

2024, A&A, 686, A212

Hothi, I., Allys, E., Semelin, B., & Boulanger, F. 2024, A&A, 686, A212

2024
[20]

A., Seiler, J., et al

Hutter, A., Watkinson, C. A., Seiler, J., et al. 2020, MNRAS, 492, 653

2020
[21]

T., Mellema, G., & Shapiro, P

Ichikawa, K., Barkana, R., Iliev, I. T., Mellema, G., & Shapiro, P. R. 2010, MN- RAS, 406, 2521

2010
[22]

& Cole, S

Lacey, C. & Cole, S. 1993, MNRAS, 262, 627

1993
[23]

R., Mondal, R., et al

Majumdar, S., Pritchard, J. R., Mondal, R., et al. 2018, MNRAS, 476, 4007

2018
[24]

McQuinn, M., Zahn, O., Zaldarriaga, M., Hernquist, L., & Furlanetto, S. R. 2006, ApJ, 653, 815

2006
[25]

T., Alvarez, M., & Shapiro, P

Mellema, G., Iliev, I. T., Alvarez, M., & Shapiro, P. R. 2006, NewA, 11, 374

2006
[26]

& Semelin, B

Meriot, R. & Semelin, B. 2024, A&A, 683, A24

2024
[27]

2024, arXiv e-prints, arXiv:2411.03093

Meriot, R., Semelin, B., & Cornu, D. 2024, arXiv e-prints, arXiv:2411.03093

2024 arXiv
[28]

2023, arXiv e-prints, arXiv:2311.03447

Mittal, S., Kulkarni, G., & Garel, T. 2023, arXiv e-prints, arXiv:2311.03447

2023 arXiv
[29]

Monaghan, J. J. 1992, ARA&A, 30, 543

1992
[30]

G., Koopmans, L

Munshi, S., Mertens, F. G., Koopmans, L. V . E., et al. 2024, A&A, 681, A62

2024
[31]

& Murray, I

Papamakarios, G. & Murray, I. 2016, arXiv e-prints, arXiv:1605.06376 Prelogovi´c, D. & Mesinger, A. 2023, MNRAS, 524, 4239

2016 arXiv
[32]

2021, MNRAS, 506, 5479 Rubiño-Martín, J

Reis, I., Fialkov, A., & Barkana, R. 2021, MNRAS, 506, 5479 Rubiño-Martín, J. A., Betancort-Rijo, J., & Patiri, S. G. 2008, MNRAS, 386, 2181

2021
[33]

2016, MNRAS, 455, 962

Semelin, B. 2016, MNRAS, 455, 962

2016
[34]

& Combes, F

Semelin, B. & Combes, F. 2002, A&A, 495, 389

2002
[35]

2007, A&A, 495, 389

Semelin, B., Combes, F., & Baek, S. 2007, A&A, 495, 389

2007
[36]

2017, MNRAS, 472, 4508

Semelin, B., Eames, E., Bolgar, F., & Caillat, M. 2017, MNRAS, 472, 4508

2017
[37]

2023, A&A, 672, A162

Semelin, B., Mériot, R., Mertens, F., et al. 2023, A&A, 672, A162

2023
[38]

2016, MNRAS, 458, 3003

Shimabukuro, H., Yoshiura, S., Takahashi, K., Yokoyama, S., & Ichiki, K. 2016, MNRAS, 458, 3003

2016
[39]

T., Subrahmanyan, R., et al

Singh, S., Jishnu, N. T., Subrahmanyan, R., et al. 2022, Nature Astronomy, 6, 607

2022
[40]

2018 [arXiv:1804.06788]

Talts, S., Betancourt, M., Simpson, D., Vehtari, A., & Gelman, A. 2018 [arXiv:1804.06788]

2018 arXiv
[41]

A., Majumdar, S., Pritchard, J

Watkinson, C. A., Majumdar, S., Pritchard, J. R., & Mondal, R. 2017, MNRAS, 472, 2436

2017
[42]

Yoshiura, S., Pindor, B., Line, J. L. B., et al. 2021, MNRAS, 505, 4775

2021
[43]

Zhao, X., Mao, Y ., Cheng, C., & Wandelt, B. D. 2022, ApJ, 926, 151

2022
[44]

Zhao, X., Mao, Y ., Zuo, S., & Wandelt, B. D. 2024, ApJ, 973, 41 Article number, page 11 of 11

2024

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.

[1] [1]

E., Alexander, P., et al

Abdurashidova, Z., Aguirre, J. E., Alexander, P., et al. 2022, ApJ, 924, 51

2022

[2] [2]

2024, MNRAS, 534, L30

Acharya, A., Mertens, F., Ciardi, B., et al. 2024, MNRAS, 534, L30

work page 2024

[3] [3]

2019, MNRAS, 488, 4440

Alsing, J., Charnock, T., Feeney, S., & Wandelt, B. 2019, MNRAS, 488, 4440

2019

[4] [4]

& Wandelt, B

Alsing, J. & Wandelt, B. 2018, MNRAS, 476, L60

work page 2018

[5] [5]

2009, A&A, 495, 389

Baek, S., Di Matteo, P., Semelin, B., Combes, F., & Revaz, Y . 2009, A&A, 495, 389

work page 2009

[6] [6]

2010, A&A, 523, A4+

Baek, S., Semelin, B., Di Matteo, P., Revaz, Y ., & Combes, F. 2010, A&A, 523, A4+

work page 2010

[7] [7]

& Hut, P

Barnes, J. & Hut, P. 1986, Nature, 324

work page 1986

[8] [8]

D., Rogers, A

Bowman, J. D., Rogers, A. E. E., Monsalve, R. A., Mozdzen, T. J., & Mahesh, N. 2018, Nature, 555, 67

2018