Extraction of the angular power spectrum produced by inflation from observations of experiments such as Simons Observatory

D. I. Novikov; K. O. Parfenov

REVIEW 3 major objections 5 minor 1 cited by

Extraction of the angular power spectrum produced by inflation from observations of experiments such as Simons Observatory

T0 review · 3 major / 5 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 fitting an analytic one-parameter distribution to histograms of CMB pixel-pair products recovers the two-point correlation function—and hence the angular power spectrum—of the inflationary Gaussian component alone…

desk verdict A novel distribution-fitting idea for CMB power spectrum extraction, but the central PDF is wrong and the identifiability claim is unsupported; desk reject. read the letter →

arxiv 2411.15959 v1 pith:42LQWDE6 submitted 2024-11-24 astro-ph.CO

D. I. Novikov , K. O. Parfenov This is my paper · ORCID

classification astro-ph.CO

keywords CosmicMicrowaveBackgroundangularpowerspectrumGaussianrandomfieldprimordialB-modepolarizationforegroundcleaningtwo-pointcorrelationfunctioninternallinearcombination

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

This paper proposes a way to pull the inflationary signal out of CMB maps without ever cleaning the map itself. The authors observe that the product of two values of a Gaussian random field at fixed angular separation has an analytic one-parameter probability distribution, and that any foreground or lensing contamination shows up only as a deviation from that distribution. They argue that fitting the analytic distribution to the observed histogram of pixel-pair products recovers the true two-point correlation of the Gaussian component, whose Legendre transform is the foreground-free angular power spectrum. If correct, this would give experiments like Simons Observatory a direct statistical route to the primordial B-mode polarization produced by inflation in the early universe.

What carries the argument

The load-bearing object is the analytic probability density function for the product of two correlated Gaussian variables separated by angle $r$: $P_{gg}(x_r, \xi_r) = \frac{1}{16\pi\sqrt{1-\gamma^2}} \int_1^\infty \frac{e^{-(t-\gamma)x_r/(4(1-\gamma^2))}}{\sqrt{t^2-1}} \, dt$, with $\gamma = \xi_r/\xi_0$. Its shape is set entirely by $\xi_r$, and its first moment equals $\xi_r$, so a fit of this one-parameter family to the observed pixel-pair product distribution is, by construction, a measurement of the Gaussian-only two-point correlation. This function converts the problem of foreground separation from map-space cleaning into a one-parameter distribution fit.

What would settle it

Simulate a sky with a known Gaussian CMB component plus realistic polarized dust and lensing, build histograms of pixel-pair products at fixed separations, fit the analytic product distribution, and compare the recovered correlation values and their Legendre power spectrum with the known input; biased recovery whenever the foreground has nonzero mean or variance comparable to the signal would falsify the central claim.

Watch

Extended reading notes

Core claim

The paper's central claim is that the two-point correlation function of the Gaussian, inflationary component of the CMB can be read off from the probability distribution of products of pixel values, without cleaning the map. For a fixed angular separation $r$, the product $x_r = s_1 s_2$ of two Gaussian values has the analytically known distribution $P_{gg}(x_r, \xi_r)$ whose only free parameter is the correlation $\xi_r$. The observed distribution is this Gaussian-product distribution plus a deviation $\Delta$ produced by foregrounds and lensing. The paper asserts that fitting $P_{gg}$ to the observed histogram, treating $\Delta$ as noise in the manner of internal linear combination, returns the true $\xi_r$, and that Legendre-decomposing these $\xi_r$ values yields the power spectrum produced by inflation alone.

Load-bearing premise

The method rests on the assumption that fitting the Gaussian product distribution to the observed pixel-pair histogram, treating all foreground and lensing deviations as noise, returns the true Gaussian correlation value without bias, and the paper states this without proof or simulation.

Editorial extensions

If this is right

For each angular separation $r$, the fitted parameter $\xi_r$ is the Gaussian-only two-point correlation, and its Legendre transform gives a foreground-cleaned angular power spectrum $C_\ell$ for temperature or polarization without constructing a clean map.
Applied to polarization, the method targets the B-mode spectrum from primordial gravitational waves, the science goal of Simons Observatory, CMB-S4, and LiteBIRD.
The statistical power of the method grows with map resolution because the number of pixel pairs at separation $r$ scales as $N \cdot r/h$, and next-generation experiments provide this large pair count.
The authors note the fit can be upgraded with modeled foreground two-point information using modified internal-linear-combination variants (cILC, MILC, LRM).

Reading between the lines

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

The paper leaves the estimation of the zero-lag variance $\xi_0$ unexamined, but the template depends on it through $\gamma = \xi_r/\xi_0$; an implementation would need to propagate map noise and foreground residuals into $\xi_0$ before $\xi_r$ can be trusted.
A direct numerical test is the natural next step: simulate a Gaussian CMB plus dust and lensing, fit $P_{gg}$ to pixel-pair histograms, and compare the recovered power spectrum with the input; this would settle whether the noise treatment of the foreground deviation holds.
The same product-distribution idea should transfer to other nearly Gaussian sky fields, such as CMB lensing convergence or cosmic infrared background maps, wherever non-Gaussian contaminants bias two-point statistics.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

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

Thanks for the report. I agree with your rejection, and the stress-test note holds up on reading the paper. The genuinely new thing here is the suggestion to fit the full PDF of pixel-pair products rather than just the mean, and to skip map cleaning entirely. That is a legitimate statistical idea: product distributions of correlated Gaussians are known, and reading the two-point function off the shape of the distribution rather than its average is a real methodological novelty within CMB analysis. If it worked, it would complement ILC and delensing. So credit for the application.

The problem is that the central derivation is not close to correct. Eq. (7) is off by a factor of two in normalization and ignores the variance ξ0. Eq. (8) is not the product PDF: for correlated unit-variance Gaussians the correct expression is (1/π√(1−γ²)) exp(γx/(1−γ²)) K₀(|x|/(1−γ²)), whereas their Eq. (8) has a different prefactor, a factor of 4 in the exponent, and no absolute value—so the integral diverges for x<0 and the distribution is not normalized. That is load-bearing: Eq. (9) claims the mean is ξr, but the object in Eq. (8) doesn't integrate to 1, so Eq. (9) doesn't follow.

Even if the PDF were correct, the identifiability argument is asserted, not proven. Eq. (6) is a tautology; Δ is simply the difference between the observed distribution and the Gaussian template, and nothing guarantees that fitting the template recovers the inflationary ξr rather than the total Gaussian correlation. If the foreground is independent Gaussian noise, the observed signal is itself Gaussian with variance ξ0+σ² and correlation ξr, so the fit will return the total Gaussian ξr—noise included. And if ξ0 is also unknown, the shape depends on two parameters, not one.

There is no simulation, no synthetic test, no error analysis, no foreground model. For a three-page methodological proposal that is fatal, not a minor gap. The citation pattern is fine; the ILC references are appropriate context.

Who is this for? Someone working on CMB component separation might read it for the novelty of the distribution-fitting idea. But it is not ready for a referee as it stands. Recommendation: desk reject, with a suggestion to the authors to fix the product PDF, include the variance parameter, and demonstrate the fitting on realistic simulations before resubmission.

Referee Report

3 major / 5 minor

Summary. This manuscript proposes a method for extracting the angular power spectrum component generated by a Gaussian inflationary field from CMB observations. The authors consider pixel-pair products x_r = s1 s2 at fixed angular separation r, derive an analytic probability distribution P_gg(x_r, ξ_r) for the case of a pure Gaussian signal, and claim that fitting this distribution to the observed distribution P(x_r) yields the correct two-point correlation ξ_r, from which the 'clean' power spectrum can be obtained. The method is asserted to separate Gaussian inflationary correlations from non-Gaussian foregrounds and lensing. No simulations or data applications are presented; the paper consists of a short derivation and a statement of intent.

Significance. The idea of using the full distribution of pixel-pair products as a statistic to separate Gaussian from non-Gaussian components is conceptually interesting and, if it worked, could provide a complementary CMB power-spectrum estimator. However, the paper's central derivation contains concrete errors: Eq. (7) is not a correctly normalized joint Gaussian PDF with the stated variance ξ0, and Eq. (8) is not a valid normalized product PDF (for γ=0 it integrates to 1/8 over positive x_r and diverges for negative x_r). Furthermore, the load-bearing identifiability assumption—that fitting P_gg(x_r, ξ_r) to the observed P(x_r) recovers the inflationary-only correlation—is stated without proof or simulation and is generically false when foregrounds include any Gaussian component. Because the method is not validated and the analytic foundation is incorrect, the paper does not establish its central claim.

major comments (3)

[Section II, Eq. (7)] Eq. (7) is not the joint probability density for two Gaussian variables with zero mean and variances ⟨g1²⟩=⟨g2²⟩=ξ0. The exponent should contain a factor 1/ξ0 inside the exponentials; as written, the density implicitly assumes ξ0=1, which contradicts the stated definitions. The prefactor is also wrong by a factor of 2: for unit variances the correct prefactor is 1/(2π√(1-γ²)), not 1/(4π√(1-γ²)). Consequently Eq. (7) is not normalized and cannot be the starting point for the derivation of Eq. (8).
[Section II, Eq. (8)] The derived product PDF P_gg(x_r) is incorrect. For the special case γ=0 (independent unit-variance Gaussians), the standard product distribution is (1/π)K_0(|x_r|). Eq. (8) instead gives, for x_r>0, (1/(16π))K_0(x_r/4), which integrates to 1/8 over the positive real line and is therefore not normalized. For x_r<0 the integral in Eq. (8) diverges because the exponent becomes positive for t>γ. Even if the prefactor and scaling are corrected, the missing absolute value in the exponent is a fatal flaw. Since the entire fitting procedure in Section II relies on this template, the claimed recovery of ξ_r via Eq. (9) is not established.
[Section II, Eqs. (5)-(6) and the paragraph 'To correctly fit...'] The central identifiability claim—that fitting P_gg(x_r, ξ_r) to the observed distribution P(x_r) gives the correct value of ξ_r—is asserted without proof, simulation, or a precision analysis. Writing P(x_r)=P_gg(x_r)+Δ(x_r) is a tautology; treating Δ as 'noise' in an ILC-like fit requires that Δ be approximately orthogonal to ∂P_gg/∂ξ_r over the fitting region, which is neither shown nor generally true. A concrete counterexample: if the foreground f is Gaussian instrumental noise of variance σ², then s=g+f is itself Gaussian with total variance ξ0+σ² and correlation ξr, so the observed product distribution is exactly P_gg(x_r; ξ0+σ², ξr). Fitting the template returns the total Gaussian correlation, which includes the noise contribution; the inflationary-only ξr is not identifiable from P(x_r) alone. The authors need to provide either a rigorous argument or end-to-end simulations demonstrating that the fit isolates the inflationary component in the presence of foregrounds.

minor comments (5)

[Title/Introduction] The title mentions Simons Observatory, but the paper contains no analysis of SO data or SO-specific forecasts; the only connection is a reference in the introduction. The scope should be clarified.
[Section II, Eq. (5)] The notation in Eq. (5) is under-specified: the integral sign is written as 'R' and the integration variables y_r, z_r are introduced without explicit limits. The Jacobian factor 1/4 is not derived in the text.
[Section II, after Eq. (4)] The statement that for a negligibly small foreground the functions P_gf and P_ff 'actually turn into delta functions' is imprecise and is not used in the subsequent derivation; the limit is not formally justified.
[Figure 1] Figure 1 shows the proposed P_gg(x_r) for three values of ξ_r/ξ0, but no comparison to a histogram or Monte Carlo sample is provided, and the vertical dashed lines are not explained in the caption. Given that Eq. (8) is the object being plotted, a direct test against a Gaussian simulation would have been instructive.
[Section III] The conclusions state that the method 'extracts the correct part of the power spectrum from observational data,' but no uncertainty quantification or validation on mock sky maps is presented. This overstates the demonstrated result.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the extraction is a parameter fit to a forward-model template, not a self-referential derivation.

full rationale

The paper's claimed derivation chain is not circular in the logical sense. It starts from a physical model (observed signal = Gaussian inflationary component + foregrounds, Eq. 1), derives analytically the product PDF Pgg for a correlated Gaussian field (Eqs. 7-8), notes that ξ_r is the mean of that PDF (Eq. 9), and then proposes to fit Pgg(x_r, ξ_r) to the observed product distribution to estimate ξ_r. This is standard forward-model parameter estimation: the template Pgg is not defined in terms of the observed P(x_r), and the resulting ξ_r is not an input repackaged as a prediction. The severe weaknesses of the paper are not circularity: the identifiability assumption that Δ in Eq. 6 can be treated as noise is unproven and could be false, and Eq. 8 appears to contain algebraic errors (incorrect coefficient, missing absolute value, divergent integral for negative x_r). These are correctness and validation failures, not circular reductions. The only self-citation, Ref. [24] by Novikov et al., appears in a closing suggestion about future ILC-like methods and is not load-bearing for the central claim. The paper is not validated against simulations or external benchmarks, but that absence of evidence is a correctness risk, not a circularity. Therefore the circularity score is 0.

Assumptions & free parameters 2 free parameters · 4 assumptions · 0 invented entities

The paper introduces no new physical entities. Its load-bearing pieces are the Gaussianity and independence assumptions, plus the unverified identifiability assumption that a Gaussian template fit recovers the inflationary correlation in the presence of unknown foregrounds. The analytic PDF itself is a derived object but is treated as an axiom-like input due to apparent errors.

free parameters (2)

ξ_r (two-point correlation function at separation r) = Fitted to observed P(x_r) for each angular bin
The method estimates the Gaussian component's two-point correlation by fitting P_gg(x_r, ξ_r) to the observed product distribution. The 'extracted' power spectrum is directly the set of fitted ξ_r values.
ξ0 (variance of the Gaussian signal) = Not specified; implicit if map is normalized
Eq. 7 and Eq. 8 lack the variance ξ0, so the PDF as written only applies to unit-variance signals. In practice ξ0 would need to be known or fitted, but the paper does not discuss how.

assumptions (4)

domain assumption Inflationary CMB fluctuations form a Gaussian random field
The entire method is premised on the Gaussian nature of the primordial perturbations. Stated in the Introduction; standard in cosmology but still an assumption.
domain assumption The Gaussian cosmological signal and foregrounds are statistically independent
Used to factor the joint probability distribution into P_gg, P_gf, and P_ff in Section II, enabling the convolution structure in Eq. 5.
ad hoc to paper Fitting P_gg to the observed distribution with the non-Gaussian part treated as noise recovers the true ξ_r
The core unproven step: the paper asserts that an ILC-like fit of the Gaussian PDF to the observed PDF extracts the correct Gaussian correlation. No derivation, simulation, or external benchmark supports this.
ad hoc to paper The analytic PDF in Eq. 8 is correct
The derivation of P_gg is not fully shown and the expression appears inconsistent with the joint PDF in Eq. 7 and with the standard result for products of correlated Gaussians.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Extraction of the angular power spectrum produced by inflation from observations of experiments such as Simons Observatory." pith.science (2026). https://pith.science/paper/42LQWDE6

@misc{pith2026241115959,
  author       = {Pith},
  title        = {Pith review of: Extraction of the angular power spectrum produced by inflation from observations of experiments such as Simons Observatory},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/42LQWDE6}},
  note         = {Machine review of arXiv:2411.15959}
}

read the original abstract

We demonstrate an approach that allows separating two-point correlations created by a Gaussian random field from correlations created by cosmic foregrounds such as polarized dust emission, gravitational lensing and other non-Gaussian signals. The result of traditional approaches should typically be a 'foreground-cleaned' two-dimensional CMB map of the anisotropy or polarization. Our method does not create a clean map, but extracts the part of the two-point correlations, or equivalently the part of the power spectrum, which is due only to the Gaussian component of the observed signal produced by inflation.

Figures

Figures reproduced from arXiv: 2411.15959 by the authors.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

Probing the Transient Far-IR Sky with PRIMA
astro-ph.IM 2025-09 accept novelty 4.0 of 10

PRIMA can expect roughly ten CMB-S4 extragalactic far-infrared transient follow-ups per year, with slow-evolving sources such as protostars and TDEs as the most feasible targets.

Reference graph

Works this paper leans on

25 extracted references · 7 canonical work pages · cited by 1 Pith paper

[1]

P. Ade, J. Aguirre, Z. Ahmed, S. Aiola, A. Ali, D. Alonso, M. A. Alvarez, K. Arnold, P. Ashton, J. Austermann, et al., Journal of Cosmology and Astroparticle Physics 2019, 056 (2019), 1808.07445

arXiv 2019
[2]

The solution to this problem inevitably faces obstacles in the form of polarized foregrounds, such as galactic dust and image distortion by gravitational lensing

and LiteBIRD [3] aim to study the CMB linear polar- ization and significantly reduce the estimation of the up- per limit for tensor-to-scalar ratio up to r ≤ 0.01 ÷ 0.001. The solution to this problem inevitably faces obstacles in the form of polarized foregrounds, such as galactic dust and image distortion by gravitational lensing. All this leads to the ...

arXiv 2024
[3]

Abazajian, G

K. Abazajian, G. E. Addison, P. Adshead, Z. Ahmed, D. Akerib, A. Ali, S. W. Allen, D. Alonso, M. Al- varez, M. A. Amin, et al., Astrophys. J 926, 54 (2022), 2008.12619

arXiv 2022
[4]

Allys, K

LiteBIRD Collaboration, E. Allys, K. Arnold, J. Au- mont, R. Aurlien, S. Azzoni, C. Baccigalupi, A. J. Ban- day, R. Banerji, R. B. Barreiro, et al., Progress of The- oretical and Experimental Physics 2023, 042F01 (2023), 2202.02773

arXiv 2023
[5]

Tegmark and G

M. Tegmark and G. Efstathiou, Mon. Not. R. Astron. Soc 281, 1297 (1996), astro-ph/9507009

arXiv 1996
[6]

An Improved Diffuse Foreground Subtraction by ILC method: CMB Map and Angular Power Spectrum using Planck and WMAP Observations

V. Sudevan, P. K. Aluri, S. K. Yadav, R. Saha, and T. Souradeep, Astrophys. J 842, 62 (2017), 1612.03401

work page Pith review arXiv 2017
[7]

Hertig, K

E. Hertig, K. Wolz, T. Namikawa, A. Baleato Lizancos, S. Azzoni, I. Abril-Cabezas, D. Alonso, C. Baccigalupi, E. Calabrese, A. Challinor, et al., Phys. Rev. D 110, 043532 (2024), 2405.01621

arXiv 2024
[8]

B. Yu, J. C. Hill, and B. D. Sherwin, Phys. Rev. D 96, 123511 (2017), 1705.02332

work page Pith review arXiv 2017

Show all 25 references

[9]

BICEP/Keck Collaboration, SPTpol Collaboration, P. A. R. Ade, Z. Ahmed, M. Amiri, A. J. Anderson, J. E. Austermann, J. S. Avva, D. Barkats, R. B. Thakur, et al., Phys. Rev. D 103, 022004 (2021), 2011.08163

2021 arXiv
[10]

Namikawa, A

T. Namikawa, A. B. Lizancos, N. Robertson, B. D. Sher- win, A. Challinor, D. Alonso, S. Azzoni, C. Baccigalupi, E. Calabrese, J. Carron, et al., Phys. Rev. D 105, 023511 (2022), 2110.09730

2022 arXiv
[11]

A. A. Starobinsky, Physics Letters B 117, 175 (1982)

1982
[12]

Matacz, Phys

A. Matacz, Phys. Rev. D 56, R1836 (1997), gr- qc/9611063

1997
[13]

C. L. Bennett, D. Larson, J. L. Weiland, N. Jarosik, G. Hinshaw, N. Odegard, K. M. Smith, R. S. Hill, B. Gold, M. Halpern, et al., Astrophys. J., Supplement 208, 20 (2013), 1212.5225

2013 arXiv
[14]

Cheng, Q.-G

C. Cheng, Q.-G. Huang, and Y.-Z. Ma, Journal of Cos- mology and Astroparticle Physics 2013, 018 (2013), 1303.4497

2013 arXiv
[15]

Planck Collaboration, P. A. R. Ade, N. Aghanim, C. Armitage-Caplan, M. Arnaud, M. Ashdown, F. Atrio- Barandela, J. Aumont, C. Baccigalupi, A. J. Banday, et al., Astron. Astrophys 571, A15 (2014), 1303.5075

2014 arXiv
[16]

Akrami, M

Planck Collaboration, Y. Akrami, M. Ashdown, J. Au- mont, C. Baccigalupi, M. Ballardini, A. J. Banday, R. B. Barreiro, N. Bartolo, S. Basak, et al., Astron. Astrophys 641, A4 (2020), 1807.06208

2020 arXiv
[17]

Akrami, M

Planck Collaboration, Y. Akrami, M. Ashdown, J. Au- mont, C. Baccigalupi, M. Ballardini, A. J. Banday, R. B. Barreiro, N. Bartolo, S. Basak, et al., Astron. Astrophys 641, A7 (2020), 1906.02552

2020 arXiv
[18]

G. B. Rybicki and W. H. Press, Astrophys. J 398, 169 (1992)

1992
[19]

Remazeilles, J

M. Remazeilles, J. Delabrouille, and J.-F. Cardoso, Mon. Not. R. Astron. Soc 418, 467 (2011), 1103.1166

2011 arXiv
[20]

Remazeilles and J

M. Remazeilles and J. Chluba, Mon. Not. R. Astron. Soc 494, 5734 (2020), 1907.00916

2020 arXiv
[21]

J. C. Hill and E. Pajer, Phys. Rev. D 88, 063526 (2013), 1303.4726

2013 arXiv
[22]

Stolyarov, M

V. Stolyarov, M. P. Hobson, A. N. Lasenby, and R. B. Barreiro, Mon. Not. R. Astron. Soc 357, 145 (2005), astro-ph/0405494

2005 arXiv
[23]

Chluba, J

J. Chluba, J. C. Hill, and M. H. Abitbol, Mon. Not. R. Astron. Soc 472, 1195 (2017), 1701.00274

2017 arXiv
[24]

Rotti and J

A. Rotti and J. Chluba, Mon. Not. R. Astron. Soc 500, 976 (2021), 2006.02458

2021 arXiv
[25]

J. P. Maillard, A. Mihalchenko, D. Novikov, A. Osipova, S. Pilipenko, and J. Silk, Phys. Rev. D 109, 023523 (2024), 2401.13415

2024 arXiv

Pith tools

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

[1] [1]

P. Ade, J. Aguirre, Z. Ahmed, S. Aiola, A. Ali, D. Alonso, M. A. Alvarez, K. Arnold, P. Ashton, J. Austermann, et al., Journal of Cosmology and Astroparticle Physics 2019, 056 (2019), 1808.07445

arXiv 2019

[2] [2]

The solution to this problem inevitably faces obstacles in the form of polarized foregrounds, such as galactic dust and image distortion by gravitational lensing

and LiteBIRD [3] aim to study the CMB linear polar- ization and significantly reduce the estimation of the up- per limit for tensor-to-scalar ratio up to r ≤ 0.01 ÷ 0.001. The solution to this problem inevitably faces obstacles in the form of polarized foregrounds, such as galactic dust and image distortion by gravitational lensing. All this leads to the ...

arXiv 2024

[3] [3]

Abazajian, G

K. Abazajian, G. E. Addison, P. Adshead, Z. Ahmed, D. Akerib, A. Ali, S. W. Allen, D. Alonso, M. Al- varez, M. A. Amin, et al., Astrophys. J 926, 54 (2022), 2008.12619

arXiv 2022

[4] [4]

Allys, K

LiteBIRD Collaboration, E. Allys, K. Arnold, J. Au- mont, R. Aurlien, S. Azzoni, C. Baccigalupi, A. J. Ban- day, R. Banerji, R. B. Barreiro, et al., Progress of The- oretical and Experimental Physics 2023, 042F01 (2023), 2202.02773

arXiv 2023

[5] [5]

Tegmark and G

M. Tegmark and G. Efstathiou, Mon. Not. R. Astron. Soc 281, 1297 (1996), astro-ph/9507009

arXiv 1996

[6] [6]

An Improved Diffuse Foreground Subtraction by ILC method: CMB Map and Angular Power Spectrum using Planck and WMAP Observations

V. Sudevan, P. K. Aluri, S. K. Yadav, R. Saha, and T. Souradeep, Astrophys. J 842, 62 (2017), 1612.03401

work page Pith review arXiv 2017

[7] [7]

Hertig, K

E. Hertig, K. Wolz, T. Namikawa, A. Baleato Lizancos, S. Azzoni, I. Abril-Cabezas, D. Alonso, C. Baccigalupi, E. Calabrese, A. Challinor, et al., Phys. Rev. D 110, 043532 (2024), 2405.01621

arXiv 2024

[8] [8]

B. Yu, J. C. Hill, and B. D. Sherwin, Phys. Rev. D 96, 123511 (2017), 1705.02332

work page Pith review arXiv 2017