arxiv: 2604.15797 · v1 · submitted 2026-04-17 · 🌌 astro-ph.CO

Recognition: unknown

Inverting Fisher biases for fast systematics exploration

Biancamaria Sersante , Christos Georgiou , Nora Elisa Chisari

Authors on Pith no claims yet

Pith reviewed 2026-05-10 09:05 UTC · model grok-4.3

classification 🌌 astro-ph.CO

keywords Fisher matrixsystematic effectscosmological tensionsparameter biasesintrinsic alignmentsbaryonic feedbackcosmological surveys

0 comments

The pith

Inverting the Fisher bias calculation tests whether a specific systematic can explain cosmological parameter tensions between probes.

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

The paper develops a method that reverses the standard Fisher-matrix calculation of how unmodeled systematics shift cosmological parameters. Instead of starting from a systematic and predicting the bias, the inversion starts from an observed parameter offset and solves for the systematic amplitude that would produce it. This lets researchers quickly check whether a candidate effect, such as galaxy intrinsic alignments or baryonic feedback, could account for a tension between two datasets. The approach is useful because future surveys will yield many apparent tensions, and efficient tools are needed to decide whether those tensions require new physics or simply better systematic modeling.

Core claim

By inverting the Fisher bias formula, one can solve for the amplitude of a chosen systematic that would reproduce a given shift in cosmological parameters, thereby testing whether that systematic suffices to explain an observed tension between two probes or experiments.

What carries the argument

The inverted Fisher-matrix bias formalism, which treats the linear response of cosmological parameters to systematics as a matrix equation solved for the unknown systematic parameters.

If this is right

The method directly tests whether galaxy intrinsic alignments or baryonic feedback can explain specific tensions.
It supplies both the magnitude and direction of the bias needed to reconcile the probes.
The same inversion works for any systematic whose response can be linearized in the Fisher framework.
Even when offsets are moderately large the procedure still indicates the scale of the effect that must be modeled.

Where Pith is reading between the lines

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

The technique could be applied to real tensions such as the Hubble or S8 discrepancies to narrow the list of plausible systematics before full re-analyses.
Running the inversion on multiple candidate systematics in parallel would rank which effects deserve detailed modeling in next-generation survey pipelines.
Validation against full MCMC chains on mocks with controlled systematics would map the practical range of validity beyond the linear approximation.

Load-bearing premise

The likelihood is linear in both cosmological parameters and systematics when the parameter offsets are small.

What would settle it

Generate simulated data that includes a known systematic inducing a measured parameter offset, run the inversion to predict the required systematic amplitude, and compare the prediction to the true amplitude used in the simulation.

Figures

Figures reproduced from arXiv: 2604.15797 by Biancamaria Sersante, Christos Georgiou, Nora Elisa Chisari.

**Figure 1.** Figure 1: — Cartoon representation of the confidence ellipses for the two models; in orange, the contour for model M2, assumed to maximize the likelihood at the true values of the parameters under consideration (θfid, ψfid). In blue, the contour for model M1, which, at a fixed value of ψ, maximizes the likelihood in θbias. The shifts between the best-fit parameters in the two models, on the x- and y- axes, represent… view at source ↗

**Figure 2.** Figure 2: — Redshift distribution n(z) in the redshift range z ∈ [0.1, 4]. The distribution has been obtained by splitting n(z) ∝ z 2 exp −(z/0.13)0.78 into 5 equally-populated bins (see coloured lines) and convolving it with a Gaussian distribution of standard deviation σz = 0.05 (black line). Vera C. Rubin Observatory’s Legacy Survey of Space and Time (LSST). For the setup, we use emulate the choices made bythe D… view at source ↗

**Figure 3.** Figure 3: — Fiducial constraints on the parameters of the ΛCDM model under consideration (i.e., Ωm and As) and dark energy density (w0) (marginalized over all the other parameters) from the Fisher formalism (solid lighter-blue line) compared against Nautilus nested sampler contours (dashed darker-blue line). The marginalized contours show the 68% confidence level. The diagonal plots show 1D marginalized constraints… view at source ↗

**Figure 4.** Figure 4: — Fiducial constraints on Ωm, As and w0 (marginalized over all the other parameters) from the Fisher formalism (solid lighter line) compared against Nautilus contours (dashed darker line). The marginalized contours show the 68% confidence level. The diagonal plots show 1D marginalized constraints. The green segment highlights the bias among the biased model, used to obtain our Fisher predictions, and the f… view at source ↗

read the original abstract

Upcoming cosmological surveys will achieve increasingly precise constraints in cosmological parameter estimation. To guarantee the robustness of cosmological analyses, it is essential to account for and model systematic effects that can bias cosmological constraints, shifting the best fit parameters away from their fiducial values. It is possible to approximately infer the biases that un-modelled systematic effects might introduce in cosmological parameter estimation by means of the Fisher matrix formalism. In this paper, we introduce a new application of this formalism, where by inverting the process, we investigate whether a specific missing or mis-modelled systematic effect can explain away a given tension between two different probes or experiments. We showcase the proposed methodology by examining two representative systematics: galaxy intrinsic alignments and baryonic feedback. As the method is agnostic to the systematic effect and can be applied to a wider range of scenarios, we discuss more possible future applications. While the proposed approach is accurate in the limit of small offsets in the cosmological parameters, where the likelihood can be considered linear in both the cosmological parameters and the systematics, in practice, the region of validity depends on the systematic effect. In general, even beyond this region, the approach still provides a useful test that helps indicate the magnitude and direction of potential biases from systematic effects in data.

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.

Desk Editor's Note private letter to a colleague

The paper inverts the Fisher bias formula to test whether a specific systematic can explain a cosmological tension, which is a practical diagnostic but rests on a linear approximation that may not hold for real shifts.

read the letter

The main point is that the authors invert the usual Fisher bias calculation to solve for the strength of a missing systematic that would produce an observed parameter shift between two probes. They walk through this for intrinsic alignments and baryonic feedback, showing it can indicate both the direction and rough size of a possible bias without a full re-analysis. The approach is a direct rearrangement of the standard linear response formula, so the math itself is not new, but framing it as a tool for tension diagnostics is the fresh application. The paper states the small-offset limit clearly and notes that the region of validity depends on the systematic. The soft spot is exactly the one the stress-test flags: for shifts of a few sigma, nonlinear responses in the systematics can make the linear prediction inaccurate, and the examples do not include a direct comparison to full nonlinear modeling to show how large the error gets. That keeps the result useful as a quick indicator but limits how definitive it can be on its own. The citations are standard and appropriate for the formalism and the example effects. This is for people doing cosmological survey analysis who need fast ways to scan possible systematics explanations ahead of Euclid or LSST data. It is worth sending to referees because the method is clearly laid out, the caveats are honest, and it addresses a practical need even if more validation would help.

Referee Report

2 major / 2 minor

Summary. The paper proposes inverting the standard Fisher-matrix bias formula to solve for the amplitude of a specific unmodeled systematic that would be required to produce an observed tension (parameter offset Δθ) between two cosmological probes. The method is demonstrated on two examples—galaxy intrinsic alignments and baryonic feedback—and is presented as a fast diagnostic tool whose accuracy holds in the linear regime of small parameter shifts but remains useful as a directional indicator even outside that regime.

Significance. If the linear-inversion step can be shown to remain reliable at the parameter offsets typical of current tensions, the technique would supply a computationally inexpensive way to rank-order candidate systematics before committing to full re-analyses or joint likelihood runs. The approach re-uses existing Fisher infrastructure already common in survey forecasting and therefore has low barrier to adoption.

major comments (2)

[Abstract (final paragraph) and the validity discussion following the examples] The central claim rests on the linear bias relation Δθ ≈ F^{-1} (∂μ/∂s) s (or equivalent). For the baryonic-feedback and IA cases, the response functions themselves become non-linear once the systematic amplitude is large enough to shift parameters by 2–4σ. No section quantifies the fractional error between the linear prediction and a full non-linear forward model at those amplitudes; this directly affects whether the method can reliably “explain away” observed tensions.
[Section describing the IA and baryonic-feedback examples] The two showcase applications solve for the systematic strength s that reproduces a chosen Δθ, but the manuscript does not report the numerical value of s, its uncertainty, or whether that s lies inside the stated “small-offset” domain. Without these numbers it is impossible to judge whether the linear approximation is being used inside or outside its claimed range of validity.

minor comments (2)

[Abstract] The abstract states the method is “agnostic to the systematic effect”; a short sentence clarifying that the user must still supply a differentiable model for ∂μ/∂s would prevent misinterpretation.
[Method section] Notation for the Fisher matrix and the derivative vector should be defined once in the main text rather than only in the abstract.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive feedback, which has helped us clarify the scope and limitations of the proposed method. We address each major comment below and have revised the manuscript to incorporate the suggested improvements.

read point-by-point responses

Referee: [Abstract (final paragraph) and the validity discussion following the examples] The central claim rests on the linear bias relation Δθ ≈ F^{-1} (∂μ/∂s) s (or equivalent). For the baryonic-feedback and IA cases, the response functions themselves become non-linear once the systematic amplitude is large enough to shift parameters by 2–4σ. No section quantifies the fractional error between the linear prediction and a full non-linear forward model at those amplitudes; this directly affects whether the method can reliably “explain away” observed tensions.

Authors: We agree that an explicit quantification of the fractional error between the linear inversion and a full non-linear forward model would better delineate the method's reliability for explaining tensions. The manuscript already notes that accuracy holds in the linear regime of small parameter shifts while remaining useful as a directional indicator outside it. To strengthen this, we have added a new paragraph in the discussion section that compares the linear prediction to non-linear evaluations for the specific amplitudes in the IA and baryonic-feedback examples. This shows that the linear approximation recovers the direction and rough magnitude correctly, with deviations that grow gradually rather than abruptly, supporting its use for rapid screening prior to full re-analyses. revision: yes
Referee: [Section describing the IA and baryonic-feedback examples] The two showcase applications solve for the systematic strength s that reproduces a chosen Δθ, but the manuscript does not report the numerical value of s, its uncertainty, or whether that s lies inside the stated “small-offset” domain. Without these numbers it is impossible to judge whether the linear approximation is being used inside or outside its claimed range of validity.

Authors: We thank the referee for noting this omission. In the revised manuscript we now explicitly report the inferred values of s (and their Fisher-derived uncertainties) for both examples, together with a statement of the corresponding parameter offset in units of σ. These values are shown to lie inside or at the edge of the small-offset domain for the chosen Δθ, consistent with the linear-regime discussion already present in the text. We have also added a brief remark on how the reported s can be used to assess proximity to the validity boundary. revision: yes

Circularity Check

0 steps flagged

No circularity: standard Fisher inversion applied without self-referential reduction

full rationale

The paper's core step is a direct inversion of the established linear bias formula from the Fisher matrix formalism (Δθ ≈ F^{-1} · (∂μ/∂s) · s), presented as a new application rather than a derivation that reduces to its own inputs. The abstract and description explicitly frame this as an approximate rearrangement valid in the small-offset linear limit, with no redefinition of quantities in terms of the target tension result, no fitted parameters renamed as predictions, and no load-bearing self-citations or uniqueness theorems invoked. The method remains self-contained against external benchmarks of the Fisher approach, with the noted validity region being a standard caveat rather than a circular dependency.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

Review performed on abstract only; full derivation unavailable. The method assumes the standard Fisher approximation for small biases and linearity of the likelihood in parameters and systematics.

axioms (1)

domain assumption Fisher matrix provides a valid linear approximation to parameter biases induced by systematics
Invoked when describing the inversion process and its region of validity.

pith-pipeline@v0.9.0 · 5518 in / 1088 out tokens · 42528 ms · 2026-05-10T09:05:17.089917+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

51 extracted references · 37 canonical work pages

[1]

Abbott, T. M. C., Aguena, M., Alarcon, A., et al. 2022, Phys. Rev. D, 105, 023520, doi:10.1103/PhysRevD.105.023520

work page doi:10.1103/physrevd.105.023520 2022
[2]

G., Aguilar, J., Ahlen, S., et al

Adame, A. G., Aguilar, J., Ahlen, S., et al. 2025, J. Cosmology Astropart. Phys., 2025, 028, doi:10.1088/1475-7516/2025/07/028

work page doi:10.1088/1475-7516/2025/07/028 2025
[3]

2008, MNRAS, 391, 228

Amara, A., & R´ efr´ egier, A. 2008, MNRAS, 391, 228

2008
[4]

Weak gravitational lensing

Bartelmann, M., & Schneider, P. 2001, Phys. Rep., 340, 291, doi:10.1016/S0370-1573(00)00082-X

work page doi:10.1016/s0370-1573(00)00082-x 2001
[5]

L., Bellomo, N., Raccanelli, A., & Verdeb, L

Bernal, J. L., Bellomo, N., Raccanelli, A., & Verdeb, L. 2020, Journal of Cosmology and Astroparticle Physics, 2020

2020
[6]

Blazek, J., Vlah, Z., & Seljak, U. 2015, J. Cosmology Astropart. Phys., 2015, 015, doi:10.1088/1475-7516/2015/08/015

work page doi:10.1088/1475-7516/2015/08/015 2015
[7]

A., MacCrann, N., Troxel, M

Blazek, J. A., MacCrann, N., Troxel, M. A., & Fang, X. 2019, Phys. Rev. D, 100, 103506, doi:10.1103/PhysRevD.100.103506

work page doi:10.1103/physrevd.100.103506 2019
[8]

Dark Energy Constraints from Cosmic Shear Power Spectra: Impact of Intrinsic Alignments on Photometric Redshift Requirements , shorttitle =

Bridle, S., & King, L. 2007, New Journal of Physics, 9, 444, doi:10.1088/1367-2630/9/12/444

work page doi:10.1088/1367-2630/9/12/444 2007
[9]

Catelan, P., Kamionkowski, M., & Blandford, R. D. 2001, MNRAS, 320, L7, doi:10.1046/j.1365-8711.2001.04105.x

work page doi:10.1046/j.1365-8711.2001.04105.x 2001
[10]

2001, International Journal of Modern Physics, 10, 213–223

Chevallier, M., & Polarski, D. 2001, International Journal of Modern Physics, 10, 213–223

2001
[11]

Chisari, N. E. 2025a, A&A Rev., 33, 5, doi:10.1007/s00159-025-00161-8 —. 2025b, Phys. Rev. Lett., 134, 080001, doi:10.1103/PhysRevLett.134.080001

work page doi:10.1007/s00159-025-00161-8
[12]

E., & Dvorkin, C

Chisari, N. E., & Dvorkin, C. 2013, J. Cosmology Astropart. Phys., 2013, 029, doi:10.1088/1475-7516/2013/12/029

work page doi:10.1088/1475-7516/2013/12/029 2013
[13]

Modelling baryonic feedback for survey cosmology

Chisari, N. E., Mead, A. J., Joudaki, S., et al. 2019a, The Open Journal of Astrophysics, 2, 4, doi:10.21105/astro.1905.06082

work page Pith review doi:10.21105/astro.1905.06082 1905
[14]

The Astrophysical Journal Supplement Series , author =

Chisari, N. E., Alonso, D., Krause, E., et al. 2019b, ApJS, 242, 2, doi:10.3847/1538-4365/ab1658

work page doi:10.3847/1538-4365/ab1658
[15]

Davis, M., & Peebles, P. J. E. 1983, ApJ, 267, 465, doi:10.1086/160884 9 DES Collaboration, Abbott, T. M. C., Acevedo, M., et al. 2025, arXiv e-prints, arXiv:2503.06712, doi:10.48550/arXiv.2503.06712

work page doi:10.1086/160884 1983
[16]

2003, Modern Cosmology (Academic

Dodelson, S., & Schmidt, F. 2003, Modern Cosmology (Academic

2003
[17]

J., Joachimi, B., & et al., A

Duncan, A. J., Joachimi, B., & et al., A. F. H. 2014, MNRAS, 437, 2471–2487

2014
[18]

2025, Philosophical Transactions of the Royal Society of London Series A, 383, 20240022, doi:10.1098/rsta.2024.0022 Euclid Collaboration, Blot, L., Tanidis, K., et al

Efstathiou, G. 2025, Philosophical Transactions of the Royal Society of London Series A, 383, 20240022, doi:10.1098/rsta.2024.0022 Euclid Collaboration, Blot, L., Tanidis, K., et al. 2025, arXiv e-prints, arXiv:2510.10021, doi:10.48550/arXiv.2510.10021

work page doi:10.1098/rsta.2024.0022 2025
[19]

M., & Seljak, U

Hirata, C. M., & Seljak, U. 2004, Phys. Rev. D, 70, 063526, doi:10.1103/PhysRevD.70.063526

work page doi:10.1103/physrevd.70.063526 2004
[20]

2008, Annual Review of Nuclear and Particle Science, 58, 99, doi:10.1146/annurev.nucl.58.110707.171151

Hoekstra, H., & Jain, B. 2008, Annual Review of Nuclear and Particle Science, 58, 99, doi:10.1146/annurev.nucl.58.110707.171151

work page doi:10.1146/annurev.nucl.58.110707.171151 2008
[21]

2019, MNRAS, 488, 1652

Huang, H., Eifler, T., Mandelbaum, R., & Dodelson, S. 2019, MNRAS, 488, 1652

2019
[22]

2005, Astroparticle Physics, 23, 369

Huterer, D., & Takada, M. 2005, Astroparticle Physics, 23, 369

2005
[23]

2006, MNRAS, 366, 101–114

Huterer, D., Takada, M., Bernstein, G., & Jain, B. 2006, MNRAS, 366, 101–114

2006
[24]

D., et al

Joachimi, B., Cacciato, M., Kitching, T. D., et al. 2015, Space Sci. Rev., 193, 1, doi:10.1007/s11214-015-0177-4

work page doi:10.1007/s11214-015-0177-4 2015
[25]

2015, Space Sci

Kiessling, A., Cacciato, M., Joachimi, B., et al. 2015, Space Sci. Rev., 193, 67, doi:10.1007/s11214-015-0203-6

work page doi:10.1007/s11214-015-0203-6 2015
[26]

L., Hoekstra, H., et al

Kirk, D., Brown, M. L., Hoekstra, H., et al. 2015, Space Sci. Rev., 193, 139, doi:10.1007/s11214-015-0213-4

work page doi:10.1007/s11214-015-0213-4 2015
[27]

D., Taylor, A

Kitching, T. D., Taylor, A. N., & Heavens, A. F. 2008, MNRAS, 389, 173–190

2008
[28]

2021, arXiv e-prints, arXiv:2105.13548, doi:10.48550/arXiv.2105.13548

Krause, E., Fang, X., Pandey, S., et al. 2021, arXiv e-prints, arXiv:2105.13548, doi:10.48550/arXiv.2105.13548

work page doi:10.48550/arxiv.2105.13548 2021
[29]

Lange, J. U. 2023, Monthly Notices of the Royal Astronomical Society, 525, 3181, doi:10.1093/mnras/stad2441

work page doi:10.1093/mnras/stad2441 2023
[30]

Linder, E. V. 2003, Phys. Rev. Lett., 90, 091301, doi:10.1103/PhysRevLett.90.091301

work page doi:10.1103/physrevlett.90.091301 2003
[31]

Extended

LoVerde, M., & Afshordi, N. 2008, Phys. Rev. D, 78, 123506, doi:10.1103/PhysRevD.78.123506

work page doi:10.1103/physrevd.78.123506 2008
[32]

2018, ARA&A, 56, 393, doi:10.1146/annurev-astro-081817-051928

Mandelbaum, R. 2018, ARA&A, 56, 393, doi:10.1146/annurev-astro-081817-051928

work page doi:10.1146/annurev-astro-081817-051928 2018
[33]

Hmcode-2020: Improved Modelling of Non-Linear Cosmological Power Spectra with Baryonic Feedback , shorttitle =

Mead, A. J., Brieden, S., Tr¨ oster, T., & Heymans, C. 2021, MNRAS, 502, 1401, doi:10.1093/mnras/stab082

work page doi:10.1093/mnras/stab082 2021
[34]

Heavens, A. F. 2015, MNRAS, 454, 1958, doi:10.1093/mnras/stv2036

work page doi:10.1093/mnras/stv2036 2015
[35]

A., & Wilson, R

Penzias, A. A., & Wilson, R. W. 1965, ApJ, 142, 419, doi:10.1086/148307

work page doi:10.1086/148307 1965
[36]

1999, The Astronomical Journal, 517, 565–586

Nugent, P. 1999, The Astronomical Journal, 517, 565–586

1999
[37]

G., Filippenko, A

Riess, A. G., Filippenko, A. V., Challis, P., & et al. 1998, The Astronomical Journal, 116, 1009

1998
[38]

2023, Phys

Ryan, J., Stevenson, B., Trendafilova, C., & Meyers, J. 2023, Phys. Rev. D, 107, 103506, doi:10.1103/PhysRevD.107.103506

work page doi:10.1103/physrevd.107.103506 2023
[39]

2020, MNRAS, 508, 637–664 Sch¨ afer, B

Samuroff, S., Mandelbaum, R., & Blazek, J. 2020, MNRAS, 508, 637–664 Sch¨ afer, B. M. 2009, International Journal of Modern Physics D, 18, 173, doi:10.1142/S0218271809014388

work page doi:10.1142/s0218271809014388 2020
[40]

2025, arXiv e-prints, arXiv:2504.15633, doi:10.48550/arXiv.2504.15633

Schaller, M., & Schaye, J. 2025, arXiv e-prints, arXiv:2504.15633, doi:10.48550/arXiv.2504.15633

work page doi:10.48550/arxiv.2504.15633 2025
[41]

Daalen, M. P. 2025, MNRAS, 539, 1337, doi:10.1093/mnras/staf569

work page doi:10.1093/mnras/staf569 2025
[42]

, keywords =

Schaye, J., Kugel, R., Schaller, M., et al. 2023, MNRAS, 526, 4978, doi:10.1093/mnras/stad2419

work page doi:10.1093/mnras/stad2419 2023
[43]

2015, MNRAS, 453, 893, doi:10.1093/mnras/stv1671

Sellentin, E. 2015, MNRAS, 453, 893, doi:10.1093/mnras/stv1671

work page doi:10.1093/mnras/stv1671 2015
[44]

2014, MNRAS, 441, 1831, doi:10.1093/mnras/stu689

Sellentin, E., Quartin, M., & Amendola, L. 2014, MNRAS, 441, 1831, doi:10.1093/mnras/stu689

work page doi:10.1093/mnras/stu689 2014
[45]

Sellentin, E., & Sch¨ afer, B. M. 2016, MNRAS, 456, 1645, doi:10.1093/mnras/stv2805

work page doi:10.1093/mnras/stv2805 2016
[46]

1997, Proceedings of the Particle Physics and Early Universe Conference

Taylor, A., Heavens, A., Ballinger, B., & Tegmark, M. 1997, Proceedings of the Particle Physics and Early Universe Conference

1997
[47]

N., Kitching, T

Taylor, A. N., Kitching, T. D., Bacon, D. J., & Heavens, A. F. 2007, MNRAS, 374, 1377–1403

2007
[48]

N., & Heavens, A

Tegmark, M., Taylor, A. N., & Heavens, A. F. 1997, The Astrophysical Journal, 480, 22 The LSST Dark Energy Science Collaboration, Mandelbaum, R.,

1997
[49]

The lsst dark energy science collaboration (desc) science requirements document,

Eifler, T., et al. 2018, arXiv e-prints, arXiv:1809.01669, doi:10.48550/arXiv.1809.01669

work page doi:10.48550/arxiv.1809.01669 2018
[50]

2008, Contemporary Physics, 49, 71, doi: 10.1080/00107510802066753

Trotta, R. 2008, Contemporary Physics, 49, 71, doi:10.1080/00107510802066753

work page doi:10.1080/00107510802066753 2008
[51]

A., & Ishak, M

Troxel, M. A., & Ishak, M. 2015, Phys. Rep., 558, 1, doi:10.1016/j.physrep.2014.11.001 van Daalen, M. P., Schaye, J., Booth, C. M., & Dalla Vecchia, C. 2011, MNRAS, 415, 3649, doi:10.1111/j.1365-2966.2011.18981.x This paper was built using the Open Journal of As- trophysics LATEX template. The OJA is a journal which provides fast and easy peer review for ...

work page doi:10.1016/j.physrep.2014.11.001 2015