pith. machine review for the scientific record. sign in

arxiv: 2605.08970 · v1 · submitted 2026-05-09 · 🌌 astro-ph.CO

Recognition: 2 theorem links

· Lean Theorem

Reionisation time field reconstruction from 21-cm Maps: Investigating predictor coherence in WDM cosmology

Dominique Aubert, \'Emilie Th\'elie, Julien Hiegel, Nicolas Mai, Rodrigo Ibata

Pith reviewed 2026-05-12 01:46 UTC · model grok-4.3

classification 🌌 astro-ph.CO
keywords reionisation21-cmCNNwarm dark matterreionisation time fieldcoherence testcosmological models
0
0 comments X

The pith

CNN predictors show coherence with CDM and high-mass WDM reionisation models but deviate for low-mass WDM.

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

This paper develops a test for the reliability of convolutional neural networks that reconstruct the reionisation time field from 21-centimeter maps. The test uses the fact that reconstructions from different redshift slices of the same map should have identical statistics if the predictor matches the true model. When the CDM model serves as reference, networks trained on 5 keV and 7 keV warm dark matter produce consistent results, much like the CDM network itself. Networks trained on 2 keV warm dark matter, however, generate reconstructions whose statistics vary with input redshift, indicating they are sensitive to mismatches between training and input models.

Core claim

By comparing the statistics of reconstructed treion(r) fields across different redshifts of the input 21-cm maps, the CNN predictors trained on CDM and on 5 and 7 keV WDM models display high-level self-consistency, whereas the 2 keV and to a lesser extent 3 keV predictors exhibit significant deviations. This demonstrates that the predictors retain sensitivity to differences in the underlying reionisation model.

What carries the argument

The redshift-coherence criterion for treion(r) reconstructions, which requires that statistical properties remain unchanged across input map redshifts when the predictor is matched to the model.

If this is right

  • CNN predictors for 5 and 7 keV WDM can be used with confidence similar to those for CDM.
  • The 2 keV WDM predictor's deviations allow exclusion of that model if observed data shows inconsistency.
  • Machine learning predictors must be validated against their training models prior to application on real observations.
  • The proposed coherence test offers a method to assess compatibility of reionisation models with 21-cm data.

Where Pith is reading between the lines

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

  • If applied to actual telescope data, the test could help determine which dark matter models are viable.
  • Similar coherence checks might improve reconstructions of other cosmic fields like the ionisation fraction.
  • Training on a range of models or using ensemble methods could mitigate inconsistencies for marginal cases like 3 keV WDM.

Load-bearing premise

The statistics of the reionisation time field reconstructions are expected to be identical for input maps at any redshift when the predictor is trained on the correct model.

What would settle it

Finding that the 2 keV WDM predictor yields treion(r) statistics that are independent of the input redshift, matching the CDM case, would falsify the reported deviations and the claim of model sensitivity.

Figures

Figures reproduced from arXiv: 2605.08970 by Dominique Aubert, \'Emilie Th\'elie, Julien Hiegel, Nicolas Mai, Rodrigo Ibata.

Figure 1
Figure 1. Figure 1: Theoretical halo mass function for the five models used in this paper at 𝑧 = 8. The vertical dotted line shows the minimal halo mass required to host star formation, computed from 𝑇vir. The bottom panel shows the fraction 𝑛wdm𝑖 (𝑚)/𝑛cdm(𝑚), where 𝑛(𝑚) = 𝑑𝑛/𝑑𝑚 and wdm𝑖 is the WDM model associated with the relevant colour. Cross markers display the fraction 𝑛wdm𝑖 (> 𝑀min)/𝑛cdm(> 𝑀min) where 𝑛(> 𝑀min) is the … view at source ↗
Figure 2
Figure 2. Figure 2: Illustration of the five DM models and their respective fields. The upper row depicts the reionisation time field, while the two rows below show 21-cm maps at redshifts 11 and 8 extracted from the region indicated by the black square in the 𝑡reion (®𝑟) maps. Each column represents a specific DM model, arranged from left to right as follows: 2 keV, 3 KeV, 5 keV, 7 keV, and CDM. These maps have been produced… view at source ↗
Figure 3
Figure 3. Figure 3: Volume fraction of neutral hydrogen 𝑄HI for the five DM models. Time goes from left to right. The upper panel describes the mean reion￾isation history. On the left, all hydrogen is neutral (=1) and on the right, all hydrogen is ionised (=0). The lower panel displays the difference 𝑄HI,WDM-𝑄HI,CDM where WDM is the corresponding model according to the colour. Shaded areas stand for the standard deviation for… view at source ↗
Figure 4
Figure 4. Figure 4: Power Spectrum 𝑃𝑘 computed from 𝑡reion (®𝑟) for the five DM models. The upper panel describes the power spectrum, both mean (solid lines) and standard deviation (shaded areas). The large-scale structures are represented on the left side of the figure, while the small scale structures are depicted on the right (respectively small and large spatial frequencies). The lower panel displays the relative differen… view at source ↗
Figure 6
Figure 6. Figure 6: Minima number count for the five DM models depicting the min￾ima of 𝑡reion (®𝑟). Time goes from left to right and the axis is standardised where 0 is the mean reionisation time for each model. The upper panel shows the mean (solid lines) and standard deviation (shaded areas) of the minima number count. The lower panel displays the relative difference 𝑁WDM − 𝑁CDM divided by 𝑁CDM where WDM is the correspondi… view at source ↗
Figure 7
Figure 7. Figure 7: Pipeline used to obtain the coherence of our predictors with respect to CDM inputs (on the left, for redshift 8 on the top and redshift 11 on the bottom). For the example, only two models is represented (CDM and WDM) and we consider the statistic 𝑄HI. First, each predictor is fed with the 21-cm CDM map at the corresponding redshift to obtain 𝑡reion (®𝑟) reconstruction. Afterward, we compute 𝑄HI for each re… view at source ↗
Figure 8
Figure 8. Figure 8: Example of predictions of 𝑡reion (®𝑟) for each predictor when the mock observations (input) is 21-cm maps from the model CDM. The left column is the 21-cm maps for 𝑧 = 11 (top) and 𝑧 = 8 (bottom). All the 𝑡reion (®𝑟) maps are predictions using a given predictor (written on the maps). Only the CDM reconstructions are shown over the full field, whereas the reconstructions for the other predictors are shown f… view at source ↗
Figure 9
Figure 9. Figure 9: Comparison of the neutral hydrogen volume fraction between redshift 11 and 8 for each model. Dashed/shaded areas stand for the standard deviation, while solid lines depict the mean. conclusion: The 2 and 3 keV predictors (𝑅 2 Δ = −4.8 and −1.2, 10 1 k [h/cMpc] 0.8 0.6 0.4 0.2 0.0 0.2 0.4 P k, z11 z8 P k, z8 2keV 3keV 5keV 7keV CDM [PITH_FULL_IMAGE:figures/full_fig_p009_9.png] view at source ↗
Figure 11
Figure 11. Figure 11: Comparison of the total isocontour length between redshift 11 and 8 for each model. Dashed/shaded areas stand for the standard deviation, while solid lines depict the mean. other models (𝑅 2 Δ = 0.87 and 0.94, for the models 5 keV and 7 keV, respectively) because their behaviours deviate from the other predictors, showing that these two predictors lost similar￾ities with the reference and can be excluded … view at source ↗
read the original abstract

The reionisation time field treion(r) captures the entire history of cosmic reionisation by mapping the moment where each region of the Universe became ionised. Previous work has shown that treion(r) can be inferred from 21-cm observations, using convolutional neural networks (CNNs). However, these CNN predictors are trained on specific reionisation models, raising critical concerns about their reliability when applied to observational data potentially differing from their training assumptions. This paper aims to propose and test a method to evaluate the coherence of our CNN predictors with respect to their input model, thereby enabling the validation or exclusion of underlying reionisation models based on their reconstruction behaviour. By setting the CDM model as reference input, we evaluate the coherence of treion(r) reconstructions by comparing them across different redshifts for several prediction models as the statistics of treion (r) reconstructions should be the same for every redshift of the input maps. Our study particularly investigates CNNs trained on cold and warm dark matter (WDM) models, with WDM particle masses of 2, 3, 5, and 7 keV. We find that the predictors trained on 5 and 7 keV WDM models exhibit high-level self-consistency similar to the CDM predictor, while the 2 keV predictor, and to a lesser extent the 3 keV predictor, display significant deviations across several metrics. These findings seem to demonstrate that CNN predictors retain sensitivity to differences in the underlying reionisation model and can be used to assess model compatibility with observations. Our results highlight the necessity of validating machine-learning predictors against their input models before applying them to real data. The method proposed here offers a pathway to more trustworthy applications of CNNs in the study of reionisation.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

2 major / 1 minor

Summary. The manuscript proposes a coherence test for CNN predictors of the reionisation time field treion(r) from 21-cm maps. Predictors are trained on CDM and WDM reionisation models (particle masses 2, 3, 5, 7 keV) and evaluated on CDM-generated 21-cm inputs by comparing cross-redshift statistics of the reconstructed treion(r) fields. The central finding is that 5 and 7 keV WDM predictors exhibit high self-consistency comparable to the CDM predictor, while the 2 keV predictor (and to a lesser extent 3 keV) shows significant deviations, interpreted as evidence that the CNNs retain sensitivity to the underlying reionisation model and can therefore be used to assess model compatibility with observations.

Significance. If the coherence diagnostic can be shown to isolate model mismatch rather than training artefacts, the approach would offer a practical route to validating ML predictors before application to real 21-cm data, which is a recognised challenge in reionisation cosmology. The systematic exploration across four WDM masses provides a graded probe of sensitivity to small-scale power suppression and constitutes a clear strength. The cross-redshift consistency criterion is a novel, observationally motivated idea that could complement existing model-selection techniques.

major comments (2)
  1. [Abstract and Methods] The coherence test is performed exclusively on CDM-generated 21-cm maps fed to all predictors (Abstract; results on cross-redshift statistics). No positive control is reported in which WDM-trained predictors are applied to matching WDM-generated maps to confirm that redshift-independent treion(r) statistics are recovered when input and training models coincide. Without this control, the deviations seen for the 2 keV (and 3 keV) predictors on CDM inputs could arise from training difficulties on more strongly suppressed small-scale power rather than from genuine model incompatibility. This directly undermines the claim that the method isolates sensitivity to reionisation-model differences.
  2. [Abstract] The coherence criterion is defined as the requirement that 'the statistics of treion(r) reconstructions should be the same for every redshift of the input maps' when CDM is used as reference (Abstract). No derivation or quantitative justification is given for why this redshift-independence must hold for a correctly matched predictor, nor is it shown that the chosen summary statistics (unspecified in the abstract) are insensitive to the input redshift under the CDM model. This assumption is load-bearing for interpreting deviations as model mismatch.
minor comments (1)
  1. [Abstract] The abstract states that 'clear differences' and 'significant deviations' are found but supplies no numerical values, error bars, number of simulations, redshifts sampled, or exact metrics used. Adding these details would allow readers to judge the magnitude and robustness of the reported effects.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their detailed and constructive report. Their comments correctly identify two areas where the current manuscript provides insufficient support for the interpretation of the coherence test. We address each point below and will revise the manuscript accordingly.

read point-by-point responses

  1. Referee: [Abstract and Methods] The coherence test is performed exclusively on CDM-generated 21-cm maps fed to all predictors (Abstract; results on cross-redshift statistics). No positive control is reported in which WDM-trained predictors are applied to matching WDM-generated maps to confirm that redshift-independent treion(r) statistics are recovered when input and training models coincide. Without this control, the deviations seen for the 2 keV (and 3 keV) predictors on CDM inputs could arise from training difficulties on more strongly suppressed small-scale power rather than from genuine model incompatibility. This directly undermines the claim that the method isolates sensitivity to reionisation-model differences.

    Authors: We agree that the absence of an explicit positive control weakens the interpretation. The CDM predictor applied to CDM maps provides an implicit demonstration of coherence, and the graded behaviour across WDM masses (high coherence at 5-7 keV, clear deviations at 2-3 keV) is consistent with model mismatch rather than a generic training artefact. Nevertheless, the referee's concern is valid. In the revised manuscript we will add the missing positive-control experiments: each WDM-trained predictor will be applied to 21-cm maps generated from its own training model, and we will report the resulting cross-redshift statistics to confirm that redshift-independent treion(r) recovery is recovered when the input and training models coincide. revision: yes

  2. Referee: [Abstract] The coherence criterion is defined as the requirement that 'the statistics of treion(r) reconstructions should be the same for every redshift of the input maps' when CDM is used as reference (Abstract). No derivation or quantitative justification is given for why this redshift-independence must hold for a correctly matched predictor, nor is it shown that the chosen summary statistics (unspecified in the abstract) are insensitive to the input redshift under the CDM model. This assumption is load-bearing for interpreting deviations as model mismatch.

    Authors: We acknowledge that the manuscript does not supply a formal derivation of the redshift-independence criterion. The physical basis is that treion(r) is a fixed, time-integrated field; under a correctly matched reionisation model the 21-cm signal at any redshift during the epoch encodes information sufficient to reconstruct the entire history, so the recovered statistics must be independent of the chosen input redshift. The summary statistics employed (power spectra and cross-redshift correlation measures, fully specified in Section 3) are chosen precisely because they are insensitive to the precise redshift slice when the model is correct. In the revised manuscript we will insert a short dedicated subsection deriving this expectation from the definition of treion(r) and demonstrating, for the CDM case, that the chosen statistics remain stable across the redshift range used. revision: yes

Circularity Check

0 steps flagged

No circularity; coherence test is an empirical comparison of output statistics

full rationale

The paper proposes an explicit coherence criterion (redshift-independence of treion(r) statistics on CDM inputs) and applies it by training separate CNNs on CDM and WDM simulations then measuring deviations in reconstructed statistics. This is a direct empirical test on held-out map properties rather than any quantity being redefined in terms of itself, a fitted parameter being relabeled as a prediction, or a central result being justified solely by self-citation. The derivation chain remains self-contained because the reported differences for 2/3 keV versus 5/7 keV models arise from observable metric mismatches, not from any equation or assumption that reduces to the input data by construction.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

The central claim rests on the assumption that reionisation histories differ measurably with WDM particle mass and that CNNs trained on one history will produce redshift-independent statistics only when the input model matches the training model. No free parameters or invented entities are mentioned in the abstract.

axioms (2)
  • domain assumption The statistics of treion(r) reconstructions must be identical across redshifts when the input cosmology matches the training model.
    Invoked to define the coherence test when CDM is used as reference.
  • domain assumption Reionisation models with different WDM particle masses (2, 3, 5, 7 keV) produce sufficiently distinct 21-cm signals and treion fields for CNNs to retain sensitivity.
    Required for the claim that deviations indicate model incompatibility.

pith-pipeline@v0.9.0 · 5643 in / 1522 out tokens · 31825 ms · 2026-05-12T01:46:14.349717+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

Reference graph

Works this paper leans on

69 extracted references · 69 canonical work pages

  1. [1]

    D., Wilkinson, M

    Alvarez, Marcelo A. and Abel, Tom , year=. Quasar H <scp>ii</scp> regions during cosmic reionization , volume=. Monthly Notices of the Royal Astronomical Society: Letters , publisher=. doi:10.1111/j.1745-3933.2007.00342.x , number=

    work page doi:10.1111/j.1745-3933.2007.00342.x 2007

  2. [2]

    , keywords =

    EMMA: an adaptive mesh refinement cosmological simulation code with radiative transfer. , keywords =. doi:10.1093/mnras/stv1896 , archivePrefix =. 1508.07888 , primaryClass =

    work page doi:10.1093/mnras/stv1896

  3. [3]

    , keywords =

    Constraints on Warm Dark Matter from Cosmological Reionization. , keywords =. doi:10.1086/322393 , archivePrefix =. astro-ph/0102304 , primaryClass =

    work page doi:10.1086/322393

  4. [4]

    , keywords =

    In the beginning: the first sources of light and the reionization of the universe. , keywords =. doi:10.1016/S0370-1573(01)00019-9 , archivePrefix =. astro-ph/0010468 , primaryClass =

    work page doi:10.1016/s0370-1573(01)00019-9

  5. [5]

    and Trac, H

    Battaglia, N. and Trac, H. and Cen, R. and Loeb, A. , year=. REIONIZATION ON LARGE SCALES. I. A PARAMETRIC MODEL CONSTRUCTED FROM RADIATION-HYDRODYNAMIC SIMULATIONS , volume=. The Astrophysical Journal , publisher=. doi:10.1088/0004-637x/776/2/81 , number=

    work page doi:10.1088/0004-637x/776/2/81

  6. [6]

    and Farahi, Arya and Cole, Shaun and Moustakas, Leonidas A

    Benson, Andrew J. and Farahi, Arya and Cole, Shaun and Moustakas, Leonidas A. and Jenkins, Adrian and Lovell, Mark and Kennedy, Rachel and Helly, John and Frenk, Carlos , year=. Dark matter halo merger histories beyond cold dark matter – I. Methods and application to warm dark matter , volume=. Monthly Notices of the Royal Astronomical Society , publisher...

    work page doi:10.1093/mnras/sts159

  7. [7]

    K., Iliev, I

    Bianco, Michele and Giri, Sambit K and Iliev, Ilian T and Mellema, Garrelt , year=. Deep learning approach for identification of H <scp>ii</scp> regions during reionization in 21-cm observations , volume=. Monthly Notices of the Royal Astronomical Society , publisher=. doi:10.1093/mnras/stab1518 , number=

    work page doi:10.1093/mnras/stab1518

  8. [8]

    K., Mellema, G., et al

    Bianco, Michele and Giri, Sambit K and Prelogović, David and Chen, Tianyue and Mertens, Florent G and Tolley, Emma and Mesinger, Andrei and Kneib, Jean-Paul , year=. Deep learning approach for identification of <scp>Hii</scp> regions during reionization in 21-cm observations – II. Foreground contamination , volume=. Monthly Notices of the Royal Astronomic...

    work page doi:10.1093/mnras/stae257

  9. [9]

    image recovery , author=

    Deep learning approach for identification of HII regions during reionization in 21-cm observations -- III. image recovery , author=. 2024 , eprint=

    work page 2024

  10. [10]

    , keywords =

    Galaxy formation by dissipationless particles heavier than neutrinos. , keywords =. doi:10.1038/299037a0 , adsurl =

    work page doi:10.1038/299037a0

  11. [11]

    , keywords =

    Halo Formation in Warm Dark Matter Models. , keywords =. doi:10.1086/321541 , archivePrefix =. astro-ph/0010389 , primaryClass =

    work page doi:10.1086/321541

  12. [12]

    , keywords =

    Toward Empirical Constraints on the Global Redshifted 21 cm Brightness Temperature During the Epoch of Reionization. , keywords =. doi:10.1086/528675 , archivePrefix =. 0710.2541 , primaryClass =

    work page doi:10.1086/528675

  13. [13]

    and Iakubovskyi, D

    Boyarsky, A. and Iakubovskyi, D. and Ruchayskiy, O. and Rudakovskyi, A. and Valkenburg, W. , year=. 21-cm observations and warm dark matter models , volume=. Physical Review D , publisher=. doi:10.1103/physrevd.100.123005 , number=

    work page doi:10.1103/physrevd.100.123005

  14. [14]

    , keywords =

    Too big to fail? The puzzling darkness of massive Milky Way subhaloes. , keywords =. doi:10.1111/j.1745-3933.2011.01074.x , archivePrefix =. 1103.0007 , primaryClass =

    work page doi:10.1111/j.1745-3933.2011.01074.x 2011

  15. [15]

    , keywords =

    A deep learning model to emulate simulations of cosmic reionization. , keywords =. doi:10.1093/mnras/stz2605 , archivePrefix =. 1905.06958 , primaryClass =

    work page doi:10.1093/mnras/stz2605 1905

  16. [16]

    2023 , eprint=

    The stability of deep learning for 21cm foreground removal across various sky models and frequency-dependent systematics , author=. 2023 , eprint=

    work page 2023

  17. [17]

    and Bolton, James S

    Choudhury, Tirthankar Roy and Puchwein, Ewald and Haehnelt, Martin G. and Bolton, James S. , year=. Lyman α emitters gone missing: evidence for late reionization? , volume=. Monthly Notices of the Royal Astronomical Society , publisher=. doi:10.1093/mnras/stv1250 , number=

    work page doi:10.1093/mnras/stv1250

  18. [18]

    , keywords =

    Hydrogen Epoch of Reionization Array (HERA). , keywords =. doi:10.1088/1538-3873/129/974/045001 , archivePrefix =. 1606.07473 , primaryClass =

    work page doi:10.1088/1538-3873/129/974/045001

  19. [19]

    de Lera Acedo et al.Nature Astronomy, 6:984–998, July 2022

    de Lera Acedo, E. and de Villiers, D. I. L. and Razavi-Ghods, N. and Handley, W. and Fialkov, A. and Magro, A. and Anstey, D. and Bevins, H. T. J. and Chiello, R. and Cumner, J. and Josaitis, A. T. and Roque, I. L. V. and Sims, P. H. and Scheutwinkel, K. H. and Alexander, P. and Bernardi, G. and Carey, S. and Cavillot, J. and Croukamp, W. and Ely, J. A. a...

    work page doi:10.1038/s41550-022-01709-9

  20. [20]

    Impact of the reduced speed of light approximation on ionization front velocities in cosmological simulations of the epoch of reionization , volume=

    Deparis, Nicolas and Aubert, Dominique and Ocvirk, Pierre and Chardin, Jonathan and Lewis, Joseph , year=. Impact of the reduced speed of light approximation on ionization front velocities in cosmological simulations of the epoch of reionization , volume=. doi:10.1051/0004-6361/201832889 , journal=

    work page doi:10.1051/0004-6361/201832889

  21. [21]

    and Cole, Shaun and Frenk, Carlos S

    Eke, Vincent R. and Cole, Shaun and Frenk, Carlos S. , year=. Cluster evolution as a diagnostic for Ω , volume=. Monthly Notices of the Royal Astronomical Society , publisher=. doi:10.1093/mnras/282.1.263 , number=

    work page doi:10.1093/mnras/282.1.263

  22. [22]

    Joint constraints on thermal relic dark matter from strong gravitational lensing, the Lyα forest, and Milky Way satellites , volume=

    Enzi, Wolfgang and Murgia, Riccardo and Newton, Oliver and Vegetti, Simona and Frenk, Carlos and Viel, Matteo and Cautun, Marius and Fassnacht, Christopher D and Auger, Matt and Despali, Giulia and McKean, John and Koopmans, Léon V E and Lovell, Mark , year=. Joint constraints on thermal relic dark matter from strong gravitational lensing, the Lyα forest,...

    work page doi:10.1093/mnras/stab1960

  23. [23]

    and Primack, Joel R

    Flores, Ricardo A. and Primack, Joel R. , year=. Observational and theoretical constraints on singular dark matter halos , volume=. doi:10.1086/187350 , journal=

    work page doi:10.1086/187350

  24. [24]

    Foidl, Horst and Rindler-Daller, Tanja and Zeilinger, Werner W. , year=. Halo formation and evolution in scalar field dark matter and cold dark matter: New insights from the fluid approach , volume=. Physical Review D , publisher=. doi:10.1103/physrevd.108.043012 , number=

    work page doi:10.1103/physrevd.108.043012

  25. [25]

    , keywords =

    Friedrich, Martina M. and Mellema, Garrelt and Alvarez, Marcelo A. and Shapiro, Paul R. and Iliev, Ilian T. , year=. Topology and sizes of HII regions during cosmic reionization: Topology and sizes of H ii regions during EoR , volume=. Monthly Notices of the Royal Astronomical Society , publisher=. doi:10.1111/j.1365-2966.2011.18219.x , number=

    work page doi:10.1111/j.1365-2966.2011.18219.x 2011

  26. [26]

    and Zaldarriaga, Matias and Hernquist, Lars , year=

    Furlanetto, Steven R. and Zaldarriaga, Matias and Hernquist, Lars , year=. The Growth of H<scp>ii</scp>Regions During Reionization , volume=. The Astrophysical Journal , publisher=. doi:10.1086/423025 , number=

    work page doi:10.1086/423025

  27. [27]

    2025 , eprint=

    Constraining Mixed Dark Matter models with high redshift Lyman-alpha forest data , author=. 2025 , eprint=

    work page 2025

  28. [28]

    Formation de galaxies pendant et après la réionisation

    Gillet, Nicolas. Formation de galaxies pendant et après la réionisation. 2016

    work page 2016

  29. [29]

    Measuring the topology of reionization with Betti numbers , volume=

    Giri, Sambit K and Mellema, Garrelt , year=. Measuring the topology of reionization with Betti numbers , volume=. Monthly Notices of the Royal Astronomical Society , publisher=. doi:10.1093/mnras/stab1320 , number=

    work page doi:10.1093/mnras/stab1320

  30. [30]

    Greig, Bradley and Mesinger, Andrei and Pober, Jonathan C. , year=. Constraints on the temperature of the intergalactic medium atz= 8.4 with 21-cm observations , volume=. Monthly Notices of the Royal Astronomical Society , publisher=. doi:10.1093/mnras/stv2618 , number=

    work page doi:10.1093/mnras/stv2618

  31. [31]

    , keywords =

    Reionisation time field reconstruction from 21 cm signal maps. , keywords =. doi:10.1051/0004-6361/202346495 , archivePrefix =. 2307.00609 , primaryClass =

    work page doi:10.1051/0004-6361/202346495

  32. [32]

    and Bolton, James S

    Iršič, Vid and Viel, Matteo and Haehnelt, Martin G. and Bolton, James S. and Molaro, Margherita and Puchwein, Ewald and Boera, Elisa and Becker, George D. and Gaikwad, Prakash and Keating, Laura C. and Kulkarni, Girish , year=. Unveiling dark matter free streaming at the smallest scales with the high redshift Lyman-alpha forest , volume=. Physical Review ...

    work page doi:10.1103/physrevd.109.043511

  33. [33]

    , keywords =

    Jenkins, A. and Frenk, C. S. and White, S. D. M. and Colberg, J. M. and Cole, S. and Evrard, A. E. and Couchman, H. M. P. and Yoshida, N. , year=. The mass function of dark matter haloes , volume=. Monthly Notices of the Royal Astronomical Society , publisher=. doi:10.1046/j.1365-8711.2001.04029.x , number=

    work page doi:10.1046/j.1365-8711.2001.04029.x 2001

  34. [34]

    The AGORA High-resolution Galaxy Simulations Comparison Project. V. Satellite Galaxy Populations in a Cosmological Zoom-in Simulation of a Milky Way Mass Halo. , keywords =. doi:10.3847/1538-4357/ad245b , archivePrefix =. 2402.05392 , primaryClass =

    work page doi:10.3847/1538-4357/ad245b

  35. [35]

    Modified initial power spectrum and too big to fail problem , volume=

    Kameli, Hamed and Baghram, Shant , year=. Modified initial power spectrum and too big to fail problem , volume=. Monthly Notices of the Royal Astronomical Society , publisher=. doi:10.1093/mnras/staa1058 , number=

    work page doi:10.1093/mnras/staa1058

  36. [36]

    2024, MNRAS, 529, 4, 3684

    Kennedy, Jacob and Carr, Jonathan Colaço and Gagnon-Hartman, Samuel and Liu, Adrian and Mirocha, Jordan and Cui, Yue , year=. Machine-learning recovery of foreground wedge-removed 21-cm light cones for high-z galaxy mapping , volume=. Monthly Notices of the Royal Astronomical Society , publisher=. doi:10.1093/mnras/stae760 , number=

    work page doi:10.1093/mnras/stae760

  37. [37]

    , keywords =

    Where Are the Missing Galactic Satellites?. , keywords =. doi:10.1086/307643 , archivePrefix =. astro-ph/9901240 , primaryClass =

    work page doi:10.1086/307643

  38. [38]

    Koopmans, Leon and Pritchard, J and Mellema, G and Aguirre, J and Ahn, K and Barkana, R and van Bemmel, I and Bernardi, G and Bonaldi, A and Briggs, F and de Bruyn, A. G. and Chang, T. C. and Chapman, E and Chen, X and Courty, B and Dayal, P. and Ferrara, A. and Fialkov, A. and Fiore, F and Ichiki, K. and Illiev, I. T. and Inoue, S and Jelic, V and Jones,...

    work page doi:10.22323/1.215.0001

  39. [39]

    Large Lyα opacity fluctuations and low CMB τ in models of late reionization with large islands of neutral hydrogen extending to z &amp;lt; 5.5 , volume=

    Kulkarni, Girish and Keating, Laura C and Haehnelt, Martin G and Bosman, Sarah E I and Puchwein, Ewald and Chardin, Jonathan and Aubert, Dominique , year=. Large Lyα opacity fluctuations and low CMB τ in models of late reionization with large islands of neutral hydrogen extending to z &amp;lt; 5.5 , volume=. Monthly Notices of the Royal Astronomical Socie...

    work page doi:10.1093/mnrasl/slz025

  40. [40]

    and Challinor, A

    Efficient Computation of Cosmic Microwave Background Anisotropies in Closed Friedmann-Robertson-Walker Models. , keywords =. doi:10.1086/309179 , archivePrefix =. astro-ph/9911177 , primaryClass =

    work page doi:10.1086/309179

  41. [41]

    Annual Review of Astronomy and Astrophysics , author =

    The Reionization of the Universe by the First Stars and Quasars. , keywords =. doi:10.1146/annurev.astro.39.1.19 , archivePrefix =. astro-ph/0010467 , primaryClass =

    work page doi:10.1146/annurev.astro.39.1.19

  42. [42]

    THE HALO MASS FUNCTION FROM EXCURSION SET THEORY

    Maggiore, Michele and Riotto, Antonio , year=. THE HALO MASS FUNCTION FROM EXCURSION SET THEORY. I. GAUSSIAN FLUCTUATIONS WITH NON-MARKOVIAN DEPENDENCE ON THE SMOOTHING SCALE , volume=. The Astrophysical Journal , publisher=. doi:10.1088/0004-637x/711/2/907 , number=

    work page doi:10.1088/0004-637x/711/2/907

  43. [43]

    2021 , eprint=

    Exploring the Cosmic Dawn with NenuFAR , author=. 2021 , eprint=

    work page 2021

  44. [44]

    2010, MNRAS, 408, 2115, doi: 10.1111/j.1365-2966.2010.17291.x

    21CMFAST: a fast, seminumerical simulation of the high-redshift 21-cm signal. , keywords =. doi:10.1111/j.1365-2966.2010.17731.x , archivePrefix =. 1003.3878 , primaryClass =

    work page doi:10.1111/j.1365-2966.2010.17731.x 2010

  45. [45]

    and Revaz, Yves and Venhola, Aku and Rejkuba, Marina and Hilker, Michael and Lutz, Katharina , year=

    Müller, Oliver and Pawlowski, Marcel S. and Revaz, Yves and Venhola, Aku and Rejkuba, Marina and Hilker, Michael and Lutz, Katharina , year=. A too-many-dwarf-galaxy-satellites problem in the M 83 group , volume=. doi:10.1051/0004-6361/202348969 , journal=

    work page doi:10.1051/0004-6361/202348969

  46. [46]

    and Mertens, F

    Munshi, S. and Mertens, F. G. and Koopmans, L. V. E. and Offringa, A. R. and Semelin, B. and Aubert, D. and Barkana, R. and Bracco, A. and Brackenhoff, S. A. and Cecconi, B. and Ceccotti, E. and Corbel, S. and Fialkov, A. and Gehlot, B. K. and Ghara, R. and Girard, J. N. and Grießmeier, J. M. and Höfer, C. and Hothi, I. and Mériot, R. and Mevius, M. and O...

    work page doi:10.1051/0004-6361/202348329

  47. [47]

    The Journal of Open Source Software , keywords =

    21cmFAST v3: A Python-integrated C code for generating 3D realizations of the cosmic 21cm signal. The Journal of Open Source Software , keywords =. doi:10.21105/joss.02582 , archivePrefix =. 2010.15121 , primaryClass =

    work page doi:10.21105/joss.02582 2010

  48. [48]

    and Birrer, Simon and Gilman, Daniel and Wechsler, Risa H

    Nadler, Ethan O. and Birrer, Simon and Gilman, Daniel and Wechsler, Risa H. and Du, Xiaolong and Benson, Andrew and Nierenberg, Anna M. and Treu, Tommaso , year=. Dark Matter Constraints from a Unified Analysis of Strong Gravitational Lenses and Milky Way Satellite Galaxies , volume=. The Astrophysical Journal , publisher=. doi:10.3847/1538-4357/abf9a3 , number=

    work page doi:10.3847/1538-4357/abf9a3

  49. [49]

    Focusing on warm dark matter with lensed high-redshift galaxies , volume=

    Pacucci, Fabio and Mesinger, Andrei and Haiman, Zoltán , year=. Focusing on warm dark matter with lensed high-redshift galaxies , volume=. Monthly Notices of the Royal Astronomical Society: Letters , publisher=. doi:10.1093/mnrasl/slt093 , number=

    work page doi:10.1093/mnrasl/slt093

  50. [50]

    and Scelfo, G

    Parimbelli, G. and Scelfo, G. and Giri, S.K. and Schneider, A. and Archidiacono, M. and Camera, S. and Viel, M. , year=. Mixed dark matter: matter power spectrum and halo mass function , volume=. Journal of Cosmology and Astroparticle Physics , publisher=. doi:10.1088/1475-7516/2021/12/044 , number=

    work page doi:10.1088/1475-7516/2021/12/044 2021

  51. [51]

    , keywords =

    The Precision Array for Probing the Epoch of Re-ionization: Eight Station Results. , keywords =. doi:10.1088/0004-6256/139/4/1468 , archivePrefix =. 0904.2334 , primaryClass =

    work page doi:10.1088/0004-6256/139/4/1468

  52. [52]

    2015 , eprint=

    U-Net: Convolutional Networks for Biomedical Image Segmentation , author=. 2015 , eprint=

    work page 2015

  53. [53]

    2022 , eprint=

    Baryonic solutions and challenges for cosmological models of dwarf galaxies , author=. 2022 , eprint=

    work page 2022

  54. [54]

    and Smith, R

    Schneider, A. and Smith, R. E. and Reed, D. , year=. Halo mass function and the free streaming scale , volume=. Monthly Notices of the Royal Astronomical Society , publisher=. doi:10.1093/mnras/stt829 , number=

    work page doi:10.1093/mnras/stt829

  55. [55]

    Constraining noncold dark matter models with the global 21-cm signal , volume=

    Schneider, Aurel , year=. Constraining noncold dark matter models with the global 21-cm signal , volume=. Physical Review D , publisher=. doi:10.1103/physrevd.98.063021 , number=

    work page doi:10.1103/physrevd.98.063021

  56. [56]

    , keywords =

    Ellipsoidal collapse and an improved model for the number and spatial distribution of dark matter haloes. , keywords =. doi:10.1046/j.1365-8711.2001.04006.x , archivePrefix =. astro-ph/9907024 , primaryClass =

    work page doi:10.1046/j.1365-8711.2001.04006.x 2001

  57. [57]

    , keywords =

    The persistent cosmic web and its filamentary structure - I. Theory and implementation. , keywords =. doi:10.1111/j.1365-2966.2011.18394.x , archivePrefix =. 1009.4015 , primaryClass =

    work page doi:10.1111/j.1365-2966.2011.18394.x 2011

  58. [58]

    , keywords =

    The persistent cosmic web and its filamentary structure - II. Illustrations. , keywords =. doi:10.1111/j.1365-2966.2011.18395.x , archivePrefix =. 1009.4014 , primaryClass =

    work page doi:10.1111/j.1365-2966.2011.18395.x 2011

  59. [59]

    2025 , eprint=

    Warm Dark Matter meets Cold Dark Matter Isocurvature , author=. 2025 , eprint=

    work page 2025

  60. [60]

    First look at the topology of reionisation redshifts in models of the epoch of reionisation , volume=

    Thélie, Emilie and Aubert, Dominique and Gillet, Nicolas and Ocvirk, Pierre , year=. First look at the topology of reionisation redshifts in models of the epoch of reionisation , volume=. doi:10.1051/0004-6361/202141748 , journal=

    work page doi:10.1051/0004-6361/202141748

  61. [61]

    , keywords =

    Topology of reionisation times: Concepts, measurements, and comparisons to Gaussian random field predictions. , keywords =. doi:10.1051/0004-6361/202244977 , archivePrefix =. 2209.11608 , primaryClass =

    work page doi:10.1051/0004-6361/202244977

  62. [62]

    , keywords =

    The Murchison Widefield Array: The Square Kilometre Array Precursor at Low Radio Frequencies. , keywords =. doi:10.1017/pasa.2012.007 , archivePrefix =. 1206.6945 , primaryClass =

    work page doi:10.1017/pasa.2012.007 2012

  63. [63]

    Imprint of Inhomogeneous Hydrogen Reionization on the Temperature Distribution of the Intergalactic Medium , volume=

    Trac, Hy and Cen, Renyue and Loeb, Abraham , year=. Imprint of Inhomogeneous Hydrogen Reionization on the Temperature Distribution of the Intergalactic Medium , volume=. The Astrophysical Journal , publisher=. doi:10.1086/595678 , number=

    work page doi:10.1086/595678

  64. [64]

    , keywords =

    AMBER: A Semi-numerical Abundance Matching Box for the Epoch of Reionization. , keywords =. doi:10.3847/1538-4357/ac5116 , archivePrefix =. 2109.10375 , primaryClass =

    work page doi:10.3847/1538-4357/ac5116

  65. [65]

    , keywords =

    LOFAR: The LOw-Frequency ARray. , keywords =. doi:10.1051/0004-6361/201220873 , archivePrefix =. 1305.3550 , primaryClass =

    work page doi:10.1051/0004-6361/201220873

  66. [66]

    http://www.w3.org/1998/Math/MathML

    Viel, Matteo and Lesgourgues, Julien and Haehnelt, Martin G. and Matarrese, Sabino and Riotto, Antonio , year=. Constraining warm dark matter candidates including sterile neutrinos and light gravitinos with WMAP and the Lyman-<mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" display="inline"><mml:mi>α</mml:mi></mml:math>forest , volume=. Physical R...

    work page doi:10.1103/physrevd.71.063534 1998

  67. [67]

    http://www.w3.org/1998/Math/MathML

    Villasenor, Bruno and Robertson, Brant and Madau, Piero and Schneider, Evan , year=. New constraints on warm dark matter from the Lyman- <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" display="inline"><mml:mrow><mml:mi>α</mml:mi></mml:mrow></mml:math> forest power spectrum , volume=. Physical Review D , publisher=. doi:10.1103/physrevd.108.02350...

    work page doi:10.1103/physrevd.108.023502 1998

  68. [68]

    Wise, John H. , year=. Cosmic reionisation , volume=. Contemporary Physics , publisher=. doi:10.1080/00107514.2019.1631548 , number=

    work page doi:10.1080/00107514.2019.1631548 2019

  69. [69]

    The Epoch of Reionization , ISBN=

    Zaroubi, Saleem , year=. The Epoch of Reionization , ISBN=. doi:10.1007/978-3-642-32362-1_2 , booktitle=

    work page doi:10.1007/978-3-642-32362-1_2