pith. sign in

arxiv: 2606.05047 · v1 · pith:NGUYZ6ZOnew · submitted 2026-06-03 · 🌌 astro-ph.CO

Full Nonlinear Velocity Reconstruction With Transformer and Ensemble Tree Machine Learning

Yulin Gong , Rachel Bean This is my paper

Pith reviewed 2026-06-28 04:39 UTC · model grok-4.3

classification 🌌 astro-ph.CO
keywords peculiar velocity reconstructionmachine learningnonlinear velocitieslarge scale structureDESI surveykSZ effectvelocity power spectrum
0
0 comments X

The pith

Machine learning models trained on simulation residuals recover nonlinear galaxy velocities more accurately than linear theory across wider scales.

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

The paper shows that both a gradient boosting decision tree and a Transformer can be trained to predict the difference between true peculiar velocities and the estimate from linear theory, using multi-scale features from galaxy positions. This residual approach yields better recovery of the velocity power spectrum and higher cross-correlation with the actual velocities over a wider range of spatial scales. Tests use mock catalogs from AbacusSummit simulations matched to DESI luminous red and emission line galaxies, including redshift uncertainty effects. The framework is then applied to cluster pairwise velocities and stacked density profiles for kSZ studies.

Core claim

Training both a gradient boosting decision tree and a Transformer to predict the residual between the actual velocity and the linear theory estimate for line-of-sight and transverse components allows the models to capture nonlinear contributions and outperform linear theory in velocity reconstruction on AbacusSummit mocks for DESI surveys.

What carries the argument

Residual prediction between true velocity and linear theory estimate, using multi-scale features fed to Transformer or gradient boosting tree models.

If this is right

  • More accurate recovery of the velocity power spectrum than linear theory alone.
  • Higher cross-correlation with true velocities maintained across a wider range of spatial scales.
  • Improved estimates of cluster pairwise velocity correlations and stacked cluster density profiles for kSZ analyses.

Where Pith is reading between the lines

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

  • The same residual-learning setup could be retrained on mocks that include different galaxy-halo prescriptions to test robustness before application to real catalogs.
  • Velocity fields reconstructed this way might tighten constraints on modified gravity or dark energy when combined with existing density-field analyses.

Load-bearing premise

The AbacusSummit mocks accurately capture the nonlinear velocity field and survey systematics that will appear in real DESI and LSST data.

What would settle it

Training the models on one set of mocks and testing on an independent simulation suite or real DESI data yields no improvement over linear theory in power spectrum recovery or cross-correlation.

Figures

Figures reproduced from arXiv: 2606.05047 by Rachel Bean, Yulin Gong.

Figure 1
Figure 1. Figure 1: FIG. 1. Illustration of the Transformer architecture used in this work. [Left] The overall workflow of our model. [Right] The detailed [PITH_FULL_IMAGE:figures/full_fig_p007_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: FIG. 2. Comparison of the reconstructed and true LOS peculiar velocities in the periodic box for the LRG sample centered at redshift z [PITH_FULL_IMAGE:figures/full_fig_p008_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: FIG. 3. Comparison of the true [blue], linear reconstructed [green], [PITH_FULL_IMAGE:figures/full_fig_p009_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: FIG. 4. Same as Fig. 2 but for lightcone samples centered at [PITH_FULL_IMAGE:figures/full_fig_p011_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: FIG. 5. Same as Fig. 3, but presents the comparison of the true [PITH_FULL_IMAGE:figures/full_fig_p012_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: FIG. 6. Comparison of the true [blue], linear-reconstructed [green] [PITH_FULL_IMAGE:figures/full_fig_p013_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: FIG. 7. Stacked kSZ temperature profiles as a function of aperture [PITH_FULL_IMAGE:figures/full_fig_p014_7.png] view at source ↗
read the original abstract

Accurate reconstruction of peculiar velocities from galaxy positions is important for probing the motion and evolution of large scale structure. They are sensitive to the cosmological effects of gravity and dark sector matter, and complement other velocity inference methods such as the kinematic Sunyaev-Zel'dovich (kSZ) imprinted in the CMB. We show that machine learning methods improve velocity reconstruction by capturing nonlinear contributions. Specifically, we train both a gradient boosting decision tree (GBDT) and a Transformer using multi-scale features to predict the residual between the actual velocity and the estimate from linear theory for both the line-of-sight and transverse components. We evaluate our approach in both periodic box and, more realistic, lightcone settings using mock galaxy catalogs from the \textsc{AbacusSummit} simulations tailored to DESI spectroscopic surveys of luminous red galaxies (LRGs) and emission line galaxies (ELGs). We also assess the impact of redshift uncertainties such as those in Rubin LSST photometry. Both models significantly outperform the linear theory with the Transformer achieving the best performance. They more accurately recover the velocity power spectrum and maintain a higher cross-correlation with the true velocities across a wider range of spatial scales. Finally, we demonstrate two applications relevant to kSZ analyses: estimating cluster pairwise velocity correlations and stacked cluster density profiles. This machine learning framework for nonlinear velocity reconstruction opens up powerful new applications of survey data from DESI, Rubin LSST, Euclid and the Roman Space Telescope.

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 / 2 minor

Summary. The paper claims that gradient boosting decision trees (GBDT) and Transformer models, trained on multi-scale features extracted from AbacusSummit mocks tailored to DESI LRG/ELG selections, can predict nonlinear velocity residuals relative to linear theory. This yields improved recovery of the velocity power spectrum and higher cross-correlation coefficients with true velocities over a wider k-range in both periodic boxes and lightcones, plus a redshift-uncertainty test for LSST; two kSZ-related applications are also shown.

Significance. If the performance gains hold under broader validation, the approach could meaningfully extend the usable k-range for peculiar-velocity analyses in DESI, LSST, Euclid and Roman data. The work supplies a concrete, simulation-grounded demonstration that nonlinear ML corrections improve upon linear theory for both line-of-sight and transverse components, which is a useful incremental step for kSZ and velocity-field cosmology.

major comments (2)
  1. [Abstract] Abstract and evaluation sections: the reported outperformance on power-spectrum and cross-correlation metrics is stated without quantitative error bars, without ablation of the multi-scale feature set, and without explicit documentation of training/validation splits or hyperparameter selection procedures. These omissions make it impossible to assess whether the claimed gains are statistically robust or sensitive to modeling choices.
  2. [Evaluation on mocks] Evaluation on mocks and applications sections: no cross-simulation validation (e.g., on a second independent simulation suite with different baryonic physics or assembly bias) is presented. Because the central motivation is application to real DESI/LSST data, the absence of any test that the reported gains survive changes in the galaxy-halo connection or unmodeled systematics is load-bearing for the transfer claim.
minor comments (2)
  1. Notation for the residual velocity components and the precise definition of the multi-scale features should be consolidated into a single table or equation block for clarity.
  2. Figure captions for the power-spectrum and cross-correlation plots should include the exact k-range over which the improvement is quantified and the number of mocks used for the error estimation.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their thoughtful comments, which have helped us identify areas to improve the clarity and robustness of our results. We address each major comment below.

read point-by-point responses

  1. Referee: [Abstract] Abstract and evaluation sections: the reported outperformance on power-spectrum and cross-correlation metrics is stated without quantitative error bars, without ablation of the multi-scale feature set, and without explicit documentation of training/validation splits or hyperparameter selection procedures. These omissions make it impossible to assess whether the claimed gains are statistically robust or sensitive to modeling choices.

    Authors: We agree with this assessment. The revised manuscript will include quantitative error bars on the power spectrum and cross-correlation metrics. We will also add an ablation study to evaluate the contribution of the multi-scale feature set and provide detailed documentation of the training/validation splits and hyperparameter optimization procedures in the methods section. revision: yes

  2. Referee: [Evaluation on mocks] Evaluation on mocks and applications sections: no cross-simulation validation (e.g., on a second independent simulation suite with different baryonic physics or assembly bias) is presented. Because the central motivation is application to real DESI/LSST data, the absence of any test that the reported gains survive changes in the galaxy-halo connection or unmodeled systematics is load-bearing for the transfer claim.

    Authors: We recognize that cross-simulation validation would provide stronger evidence for the applicability to real data. Our work utilizes the AbacusSummit simulations, and conducting validation on additional independent suites is beyond the scope of the current study due to computational constraints. In the revision, we will add a discussion of this limitation in the conclusions and suggest it as an important direction for future research. revision: partial

Circularity Check

0 steps flagged

No significant circularity; performance measured against external simulation truth

full rationale

The paper trains GBDT and Transformer models on multi-scale features from AbacusSummit mocks to predict velocity residuals relative to linear theory, then reports power-spectrum recovery and cross-correlation on held-out test mocks (both periodic boxes and lightcones) plus a redshift-uncertainty test. These metrics are computed against the simulation's independent ground-truth velocities, not by construction from the fitted model itself. No self-citation chain, ansatz smuggling, or renaming of known results forms the load-bearing step for the central empirical claims. The derivation is therefore self-contained against the stated simulation benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract-only review; no explicit free parameters, axioms, or invented entities are stated. The approach implicitly assumes that simulation truth labels are sufficient to train a generalizable nonlinear correction and that multi-scale features capture the relevant physics.

pith-pipeline@v0.9.1-grok · 5783 in / 1232 out tokens · 24294 ms · 2026-06-28T04:39:50.647181+00:00 · methodology

discussion (0)

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

Reference graph

Works this paper leans on

87 extracted references · 5 canonical work pages · 2 internal anchors

  1. [1]

    Pairwise Velocity Statistic The pairwise velocity is a useful summary statistic of the cosmic velocity field, characterizing the average relative in- fall of tracer pairs as a function of their comoving separation. Following [73], the pairwise velocity estimator in a separation bin centered atris, V(r)= P i<j v∥,i −v ∥,j ci j P i<j c 2 i j .(15) The sum i...

  2. [2]

    displacement-only

    Stacked kSZ profile We also consider the application of the reconstructed ve- locities to estimate a kSZ-derived halo electron density pro- file, a useful probe of the spatial distribution of ionized gas in galaxy clusters. This is obtained by stacking the kSZ temper- atures centered on cluster halos after weighting by their LOS velocities. In our analysi...

  3. [3]

    Simulated kSZ Map To assess how velocity reconstruction errors propagate into a practical kSZ analysis, we construct a mock kSZ temper- ature map from the halo lightcone catalog, following [80]. In the non-relativistic limit, the kSZ temperature fluctuation along the line of sight ˆncan be written as ∆TkS Z(ˆn) T0 =− σT c Z ne vlos dl,(A1) whereσ T is the...

  4. [4]

    Pairwise kSZ Momentum and Optical Depth Fitting The kSZ pairwise momentum is given by PkS Z(r,z i)=− P i j(∆ ¯TAP,i −∆ ¯TAP,j)ci j P i j c2 i j ,(A11) where the sum is over all pairs of tracers with separations in a bin centered onr, and in a redshift bin centered onz.∆ ¯TAP,i is the kSZ temperature obtained from simulated maps by using aperture photometr...

  5. [5]

    Kaiser, Mon

    N. Kaiser, Mon. Not. Roy. Astron. Soc.227, 1 (1987)

    1987

  6. [6]

    M. A. Strauss and J. A. Willick, Phys. Rep.261, 271 (1995), arXiv:astro-ph/9502079 [astro-ph]

    Pith/arXiv arXiv 1995

  7. [7]

    Howlett, K

    C. Howlett, K. Said, J. R. Lucey, M. Colless, F. Qin, Y . Lai, R. B. Tully, and T. M. Davis, Mon. Not. Roy. Astron. Soc.515, 953 (2022), arXiv:2201.03112 [astro-ph.CO]

    arXiv 2022

  8. [8]

    W. J. Percival and M. White, Mon. Not. Roy. Astron. Soc.393, 297 (2009), arXiv:0808.0003 [astro-ph]

    Pith/arXiv arXiv 2009

  9. [9]

    M. J. Hudson and S. J. Turnbull, Astrophys. J. Lett.751, L30 (2013), arXiv:1203.4814 [astro-ph.CO]

    Pith/arXiv arXiv 2013

  10. [10]

    K. Said, M. Colless, C. Magoulas, J. R. Lucey, and M. J. Hudson, Mon. Not. Roy. Astron. Soc.497, 1275 (2020), arXiv:2007.04993 [astro-ph.CO]

    arXiv 2020

  11. [11]

    R. J. Turner, (2024), arXiv:2411.19484 [astro-ph.CO]

    arXiv 2024

  12. [12]

    R. A. Sunyaev and Y . B. Zeldovich, Mon. Not. Roy. Astron. Soc.190, 413 (1980)

    1980

  13. [13]

    Hand et al., Phys

    N. Hand et al., Phys. Rev. Lett.109, 041101 (2012), arXiv:1203.4219 [astro-ph.CO]

    Pith/arXiv arXiv 2012

  14. [14]

    Schaan et al

    E. Schaan et al. (Atacama Cosmology Telescope), Phys. Rev. D 103, 063513 (2021), arXiv:2009.05557 [astro-ph.CO]

    arXiv 2021

  15. [15]

    Laguë, M

    A. Laguë, M. S. Madhavacheril, K. M. Smith, S. Fer- raro, and E. Schaan, Phys. Rev. Lett.134, 151003 (2025), arXiv:2411.08240 [astro-ph.CO]

    arXiv 2025

  16. [16]

    Gong et al., Phys

    Y . Gong et al., Phys. Rev. D113, 063538 (2026), arXiv:2511.23417 [astro-ph.CO]

    arXiv 2026

  17. [17]

    C. M. Springob, C. Magoulas, M. Colless, J. Mould, P. Er- dogdu, D. H. Jones, J. R. Lucey, L. Campbell, and C. J. Fluke, Mon. Not. Roy. Astron. Soc.445, 2677 (2014), arXiv:1409.6161 [astro-ph.CO]

    Pith/arXiv arXiv 2014

  18. [18]

    J. S. Bullock and M. Boylan-Kolchin, Ann. Rev. Astron. Astro- phys.55, 343 (2017), arXiv:1707.04256 [astro-ph.CO]

    arXiv 2017

  19. [19]

    Battaglia, S

    N. Battaglia, S. Ferraro, E. Schaan, and D. Spergel, JCAP11, 040, arXiv:1705.05881 [astro-ph.CO]

    Pith/arXiv arXiv

  20. [20]

    D. J. Eisenstein, H.-j. Seo, and M. J. White, Astrophys. J.664, 660 (2007), arXiv:astro-ph/0604361

    Pith/arXiv arXiv 2007

  21. [21]

    D. J. Eisenstein, H.-j. Seo, E. Sirko, and D. Spergel, Astrophys. J.664, 675 (2007), arXiv:astro-ph/0604362

    Pith/arXiv arXiv 2007

  22. [22]

    Padmanabhan, X

    N. Padmanabhan, X. Xu, D. J. Eisenstein, R. Scalzo, A. J. Cuesta, K. T. Mehta, and E. Kazin, Mon. Not. Roy. Astron. Soc. 427, 2132 (2012), arXiv:1202.0090 [astro-ph.CO]

    Pith/arXiv arXiv 2012

  23. [23]

    X. Xu, A. J. Cuesta, N. Padmanabhan, D. J. Eisenstein, and C. K. McBride, Mon. Not. Roy. Astron. Soc.431, 2834 (2013), arXiv:1206.6732 [astro-ph.CO]

    Pith/arXiv arXiv 2013

  24. [24]

    Anderson et al

    L. Anderson et al. (BOSS), Mon. Not. Roy. Astron. Soc.439, 83 (2014), arXiv:1303.4666 [astro-ph.CO]

    Pith/arXiv arXiv 2014

  25. [25]

    Gil-Marin et al

    H. Gil-Marin et al. (eBOSS), Mon. Not. Roy. Astron. Soc.498, 2492 (2020), arXiv:2007.08994 [astro-ph.CO]

    arXiv 2020

  26. [26]

    K. C. Chan, G. Lu, and X. Wang, Mon. Not. Roy. Astron. Soc. 529, 1667 (2024), arXiv:2311.12611 [astro-ph.CO]

    arXiv 2024

  27. [27]

    White, Mon

    M. White, Mon. Not. Roy. Astron. Soc.450, 3822 (2015), arXiv:1504.03677 [astro-ph.CO]

    Pith/arXiv arXiv 2015

  28. [28]

    S.-F. Chen, Z. Vlah, and M. White, JCAP09, 017, arXiv:1907.00043 [astro-ph.CO]

    arXiv 1907

  29. [29]

    Schaan et al

    E. Schaan et al. (ACTPol), Phys. Rev. D93, 082002 (2016), arXiv:1510.06442 [astro-ph.CO]

    Pith/arXiv arXiv 2016

  30. [30]

    Zhang, J

    P. Zhang, J. Pan, and Y . Zheng, Physical Review D87, 10.1103/physrevd.87.063526 (2013)

    work page doi:10.1103/physrevd.87.063526 2013

  31. [31]

    Blake and R

    C. Blake and R. J. Turner 10.33232/001c.124509 (2024), arXiv:2408.05660 [astro-ph.CO]

    work page doi:10.33232/001c.124509 2024

  32. [32]

    The DESI DR1 Peculiar Velocity Survey: growth rate measurements from the maximum likelihood fields method

    Y . Lai et al. 10.5281/zenodo.17602818 (2025), arXiv:2512.03229 [astro-ph.CO]

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.5281/zenodo.17602818 2025

  33. [33]

    Aghamousa et al

    A. Aghamousa et al. (DESI), (2016), arXiv:1611.00036 [astro- ph.IM]

    Pith/arXiv arXiv 2016

  34. [34]

    LSST Science Collaboration, LSST Science Book, version 2.0 (2009), arXiv:0912.0201 [astro-ph.IM]

    Pith/arXiv arXiv 2009

  35. [35]

    Mellier et al

    Y . Mellier et al. (Euclid), Astron. Astrophys.697, A1 (2025), arXiv:2405.13491 [astro-ph.CO]

    arXiv 2025

  36. [36]

    Spergel et al., (2015), arXiv:1503.03757 [astro-ph.IM]

    D. Spergel et al., (2015), arXiv:1503.03757 [astro-ph.IM]

    Pith/arXiv arXiv 2015

  37. [37]

    B. R. Guachalla, E. Schaan, B. Hadzhiyska, and S. Ferraro, Phys. Rev. D109, 103533 (2024), arXiv:2312.12435 [astro- ph.CO]

    arXiv 2024

  38. [38]

    Hadzhiyska, S

    B. Hadzhiyska, S. Ferraro, B. R. Guachalla, and E. Schaan, Phys. Rev. D109, 103534 (2024), arXiv:2312.12434 [astro- ph.CO]

    arXiv 2024

  39. [39]

    Y . B. Zel’dovich, Astron. Astrophys.5, 84 (1970)

    1970

  40. [40]

    X. Xu, N. Padmanabhan, D. J. Eisenstein, K. T. Mehta, and A. J. Cuesta, Mon. Not. Roy. Astron. Soc.427, 2146 (2012), arXiv:1202.0091 [astro-ph.CO]

    Pith/arXiv arXiv 2012

  41. [41]

    Ma, G.-D

    Y .-Z. Ma, G.-D. Gong, N. Sui, and P. He, Mon. Not. Roy. As- tron. Soc.475, 379 (2018), arXiv:1711.08756 [astro-ph.CO]

    Pith/arXiv arXiv 2018

  42. [42]

    R. J. Turner and C. Blake, Mon. Not. Roy. Astron. Soc.526, 337 (2023), arXiv:2306.16664 [astro-ph.CO]

    arXiv 2023

  43. [43]

    A. J. Tishue, C. Shiveshwarkar, and G. Holder, (2025), arXiv:2510.25821 [astro-ph.CO]

    arXiv 2025

  44. [44]

    Suresh, H

    R. Suresh, H. Bishnoi, A. V . Kuklin, A. Parikh, M. Molokeev, R. Harinarayanan, S. Gharat, and P. Hiba, Front. in Phys.12, 1322162 (2024)

    2024

  45. [45]

    Gong and R

    Y . Gong and R. Bean, Phys. Rev. D112, 023527 (2025), arXiv:2505.12720 [astro-ph.CO]

    arXiv 2025

  46. [46]

    Ronneberger, P

    O. Ronneberger, P. Fischer, and T. Brox, arXiv e-prints , arXiv:1505.04597 (2015), arXiv:1505.04597 [cs.CV]

    Pith/arXiv arXiv 2015

  47. [47]

    Wang and X

    Y . Wang and X. Yang, Astrophys. J.969, 76 (2024), arXiv:2406.14101 [astro-ph.IM]

    arXiv 2024

  48. [48]

    X. Xiao, J. Ding, X. Luo, S. K. Lan, L. Xiao, S. Liu, X. Wang, L. Zhang, and X.-D. Li, Astrophys. J.994, 204 (2025), arXiv:2411.11280 [astro-ph.CO]

    arXiv 2025

  49. [49]

    Dosovitskiy, L

    A. Dosovitskiy, L. Beyer, A. Kolesnikov, D. Weissenborn, X. Zhai, T. Unterthiner, M. Dehghani, M. Minderer, G. Heigold, S. Gelly, J. Uszkoreit, and N. Houlsby, arXiv e-prints , arXiv:2010.11929 (2020), arXiv:2010.11929 [cs.CV]

    Pith/arXiv arXiv 2010

  50. [50]

    Z. Liu, Y . Lin, Y . Cao, H. Hu, Y . Wei, Z. Zhang, S. Lin, and B. Guo, arXiv e-prints , arXiv:2103.14030 (2021), arXiv:2103.14030 [cs.CV]

    Pith/arXiv arXiv 2021

  51. [51]

    C. Li, Y . Zhang, C. Cui, D. Fan, Y . Zhao, X.-B. Wu, B. He, Y . Xu, S. Li, J. Han, Y . Tao, L. Mi, H. Yang, and S. Yang, Monthly Notices of the Royal Astronomical Society 10.1093/mnras/stab1650 (2021)

    work page doi:10.1093/mnras/stab1650 2021

  52. [52]

    Sahakyan, V

    N. Sahakyan, V . Vardanyan, and M. Khachatryan, Mon. Not. Roy. Astron. Soc.519, 3000 (2022), arXiv:2212.06614 [astro- ph.HE]

    arXiv 2022

  53. [53]

    Tolamatti, K

    A. Tolamatti, K. K. Singh, and K. K. Yadav, Mon. Not. Roy. As- tron. Soc.523, 5341 (2023), arXiv:2306.15393 [astro-ph.HE]

    arXiv 2023

  54. [54]

    Coronado-Blázquez, Mon

    J. Coronado-Blázquez, Mon. Not. Roy. Astron. Soc.521, 4156 (2023), arXiv:2303.07922 [astro-ph.HE]

    arXiv 2023

  55. [55]

    Donoso-Oliva, I

    C. Donoso-Oliva, I. Becker, P. Protopapas, G. Cabrera-Vives, M. Vishnu, and H. Vardhan, Astron. Astrophys.670, A54 (2023)

    2023

  56. [56]

    H. W. Leung and J. Bovy, Monthly Notices of the Royal Astro- nomical Society527, 1494–1520 (2023)

    2023

  57. [57]

    Gong and R

    Y . Gong and R. Bean, Phys. Rev. D109, 123525 (2024), arXiv:2403.04664 [astro-ph.CO]

    arXiv 2024

  58. [58]

    Signor et al

    T. Signor et al. (Euclid), (2024), arXiv:2402.04800 [astro- 19 ph.CO]

    arXiv 2024

  59. [59]

    J. H. Friedman, The Annals of Statistics29, 1189 (2001)

    2001

  60. [60]

    G. Ke, Q. Meng, T. Finley, T. Wang, W. Chen, W. Ma, Q. Ye, and T.-Y . Liu, in Advances in Neural Information Processing Systems, V ol. 30 (2017)

    2017

  61. [61]

    Vaswani, N

    A. Vaswani, N. Shazeer, N. Parmar, J. Uszkoreit, L. Jones, A. N. Gomez, Ł. Kaiser, and I. Polosukhin, in Advances in Neural Information Processing Systems, V ol. 30 (2017)

    2017

  62. [62]

    N. A. Maksimova, L. H. Garrison, D. J. Eisenstein, B. Hadzhiyska, S. Bose, and T. P. Satterthwaite, Mon. Not. Roy. Astron. Soc.508, 4017 (2021), arXiv:2110.11398 [astro- ph.CO]

    arXiv 2021

  63. [63]

    L. H. Garrison, D. J. Eisenstein, and P. A. Pinto, Mon. Not. Roy. Astron. Soc.485, 3370 (2019), arXiv:1810.02916 [astro- ph.CO]

    Pith/arXiv arXiv 2019

  64. [64]

    L. H. Garrison, D. J. Eisenstein, D. Ferrer, N. A. Maksimova, and P. A. Pinto, Mon. Not. Roy. Astron. Soc.508, 575 (2021), arXiv:2110.11392 [astro-ph.CO]

    arXiv 2021

  65. [65]

    Aghanim et al

    N. Aghanim et al. (Planck), Astron. Astrophys.641, A6 (2020), [Erratum: Astron.Astrophys. 652, C4 (2021)], arXiv:1807.06209 [astro-ph.CO]

    Pith/arXiv arXiv 2020

  66. [66]

    S. Yuan, L. H. Garrison, B. Hadzhiyska, S. Bose, and D. J. Eisenstein, Mon. Not. Roy. Astron. Soc.510, 3301 (2022), arXiv:2110.11412 [astro-ph.CO]

    arXiv 2022

  67. [67]

    Yuan et al., Mon

    S. Yuan et al., Mon. Not. Roy. Astron. Soc.530, 947 (2024), arXiv:2306.06314 [astro-ph.CO]

    arXiv 2024

  68. [68]

    Zheng, A

    Z. Zheng, A. L. Coil, and I. Zehavi, Astrophys. J.667, 760 (2007), arXiv:astro-ph/0703457

    Pith/arXiv arXiv 2007

  69. [69]

    Hadzhiyska et al., (2023), arXiv:2305.11935 [astro-ph.CO]

    B. Hadzhiyska et al., (2023), arXiv:2305.11935 [astro-ph.CO]

    arXiv 2023

  70. [70]

    J. L. Tinker, B. E. Robertson, A. V . Kravtsov, A. Klypin, M. S. Warren, G. Yepes, and S. Gottlöber, Astrophys. J.724, 878 (2010), arXiv:1001.3162 [astro-ph.CO]

    Pith/arXiv arXiv 2010

  71. [71]

    Hadzhiyska, L

    B. Hadzhiyska, L. H. Garrison, D. Eisenstein, and S. Bose, Mon. Not. Roy. Astron. Soc.509, 2194 (2021), arXiv:2110.11413 [astro-ph.CO]

    arXiv 2021

  72. [72]

    Desjacques, D

    V . Desjacques, D. Jeong, and F. Schmidt, Phys. Rept.733, 1 (2018), arXiv:1611.09787 [astro-ph.CO]

    Pith/arXiv arXiv 2018

  73. [73]

    S. C. Hotinli, K. M. Smith, M. S. Madhavacheril, and M. Kamionkowski, Phys. Rev. D104, 083529 (2021), arXiv:2108.02207 [astro-ph.CO]

    arXiv 2021

  74. [74]

    S. C. Hotinli, E. Pierpaoli, S. Ferraro, and K. Smith, Phys. Rev. D108, 083508 (2023), arXiv:2305.15462 [astro-ph.CO]

    arXiv 2023

  75. [75]

    Soergel, A

    B. Soergel, A. Saro, T. Giannantonio, G. Efstathiou, and K. Dolag, Mon. Not. Roy. Astron. Soc.478, 5320 (2018), arXiv:1712.05714 [astro-ph.CO]

    Pith/arXiv arXiv 2018

  76. [76]

    Hadzhiyska, S

    B. Hadzhiyska, S. Ferraro, R. Pakmor, S. Bose, A. M. Delgado, C. Hernández-Aguayo, R. Kannan, V . Springel, S. D. M. White, and L. Hernquist, Mon. Not. Roy. Astron. Soc.526, 369 (2023), arXiv:2305.00992 [astro-ph.CO]

    arXiv 2023

  77. [77]

    P. G. Ferreira, R. Juszkiewicz, H. A. Feldman, M. Davis, and A. H. Jaffe, Astrophys. J. Lett.515, L1 (1999), arXiv:astro- ph/9812456

    arXiv 1999

  78. [78]

    J. E. Forero-Romero, Y . Hoffman, S. Gottloeber, A. Klypin, and G. Yepes, Mon. Not. Roy. Astron. Soc.396, 1815 (2009), arXiv:0809.4135 [astro-ph]

    Pith/arXiv arXiv 2009

  79. [79]

    Lucie-Smith, H

    L. Lucie-Smith, H. V . Peiris, and A. Pontzen, Mon. Not. Roy. Astron. Soc.490, 331 (2019), arXiv:1906.06339 [astro-ph.CO]

    arXiv 2019

  80. [80]

    Wanget al., Mon

    Y . Wanget al., Mon. Not. Roy. Astron. Soc.409, 737 (2010), arXiv:1006.3517 [astro-ph.CO]

    Pith/arXiv arXiv 2010

Showing first 80 references.