Publications

Large-scale structure

8 papers

Reconstructed linear matter power spectrum divided by the fiducial prediction, with 68% and 95% Monte Carlo bands.
Reconstructed linear matter power spectrum divided by the fiducial prediction, with 68% and 95% Monte Carlo bands.

2026 · arXiv preprint

Finding the distribution of matter using lenses – I: deconvolution-based reconstruction with CMB lensing

Dawn, A., Jiang, J.-Q., Hazra, D. K., L'Huillier, B., Shafieloo, A.

We reconstruct the linear matter power spectrum at z=0z = 0 from joint Planck PR4, ACT DR6 and SPT-3G CMB lensing, using a modified Richardson–Lucy deconvolution. The result follows the linear prediction on large scales, lies systematically higher for k≳0.1 Mpc−1k \gtrsim 0.1\,\mathrm{Mpc}^{-1}, and partly preserves the BAO feature.

arXiv:2609.08457

Large-scale structureGravitational lensing

Reconstructions of a 5% oscillatory feature from 1000 noisy LSST-like mocks (blue), their median (green) and the truth (red).
Reconstructions of a 5% oscillatory feature from 1000 noisy LSST-like mocks (blue), their median (green) and the truth (red).

2026 · arXiv preprint

Finding the distribution of matter using lenses – II: deconvolution-based reconstruction with 3×23\times2pt measurements

Jiang, J.-Q., Dawn, A., Hazra, D. K., L'Huillier, B., Shafieloo, A.

A regularised Richardson–Lucy framework tests scale-dependent departures from the nonlinear matter power spectrum using galaxy clustering, galaxy–galaxy lensing and cosmic shear. On LSST Year-10-like mocks, oscillations of about 1% or more are recovered over 0.1≲k≲0.5 Mpc−10.1 \lesssim k \lesssim 0.5\,\mathrm{Mpc}^{-1}, and a 1% oscillation is detected at about 2.6σ2.6\sigma.

arXiv:2609.08460

Large-scale structureGravitational lensing

Ratio mu(k,z) of the k-essence to ΛCDM gravitational potential for sound speed cs2 = 10-4, non-linear (points) versus linear (dashed).
Ratio μ(k,z)\mu(k,z) of the k-essence to ΛCDM gravitational potential for sound speed cs2=10−4c_s^2 = 10^{-4}, non-linear (points) versus linear (dashed).

2020 · JCAP 04 (2020) 039

Parametrising non-linear dark energy perturbations

Hassani, F., L'Huillier, B., Shafieloo, A., Kunz, M., Adamek, J.

Using k-evolution N-body simulations, we quantify non-linear k-essence dark energy perturbations through an effective modification μ\mu of the Poisson equation, show that linear theory is accurate at large sound speeds, and propose a simulation-calibrated parametrisation of μ\mu.

DOI · arXiv:1910.01105

Dark energyLarge-scale structureSimulations

Sub-halo counts in a 60h-1Mpc shell of the Horizon Run 4 lightcone at z = 0.36, projected on the sphere.
Sub-halo counts in a 60 h−1 Mpc60\,h^{-1}\,\mathrm{Mpc} shell of the Horizon Run 4 lightcone at z=0.36z = 0.36, projected on the sphere.

2018 · MNRAS 477 (2018) 2772

Cylinders out of a top hat: counts-in-cells for projected densities

Uhlemann, C., Pichon, C., Codis, S., L'Huillier, B., Kim, J., et al.

Large-deviation statistics predict the one-point PDF and clustering of projected densities in cylinders, in agreement with the Horizon Run 4 simulation to within a few percent in the quasi-linear regime: a tool for photometric surveys such as DES and Euclid.

DOI · arXiv:1711.04767

Large-scale structureSimulations

Primordial power spectra with features (left) and the resulting linear matter power spectra relative to Planck 2015 (right).
Primordial power spectra with features (left) and the resulting linear matter power spectra relative to Planck 2015 (right).

2018 · MNRAS 477 (2018) 2503

Probing features in the primordial perturbation spectrum with large-scale structure data

L'Huillier, B., Shafieloo, A., Hazra, D. K., Smoot, G. F., Starobinsky, A. A.

Can large-scale structure distinguish primordial power spectra with features that the CMB cannot? With 15 DESI-like N-body simulations, we show that the halo mass function and two-point correlation function cannot, but simple counts-in-cells statistics can.

DOI · arXiv:1710.10987

Early UniverseLarge-scale structureSimulationsFirst author

Halo density against dark matter density in spheres of 15h-1Mpc at z = 0, with the reconstructed bias function (red).
Halo density against dark matter density in spheres of 15 h−1 Mpc15\,h^{-1}\,\mathrm{Mpc} at z=0z = 0, with the reconstructed bias function (red).

2018 · MNRAS 473 (2018) 5098

A question of separation: disentangling tracer bias and gravitational non-linearity with counts-in-cells statistics

Uhlemann, C., Feix, M., Codis, S., Pichon, C., Bernardeau, F., et al. (incl. L'Huillier, B.)

We model the bias of tracer densities in spheres with a quadratic bias relation and the large-deviation dark matter PDF, validated on Horizon Run 4 subhaloes. This allows a joint estimate of the non-linear dark matter variance and the bias parameters.

DOI · arXiv:1705.08901

Large-scale structureSimulations

Density-dependent clustering bias b(rho) in Horizon Run 4 at z = 0.7 (points) against the saddle-point prediction (lines).
Density-dependent clustering bias b(ρ)b(\rho) in Horizon Run 4 at z=0.7z = 0.7 (points) against the saddle-point prediction (lines).

2017 · MNRAS 466 (2017) 2067

Beyond Kaiser bias: mildly non-linear two-point statistics of densities in distant spheres

Uhlemann, C., Codis, S., Kim, J., Pichon, C., Bernardeau, F., et al. (incl. L'Huillier, B.)

Parameter-free analytic bias functions extend Kaiser bias into the mildly non-linear regime, using large-deviation statistics and spherical collapse. They match Horizon Run 4 at the percent level down to about 10 h−1 Mpc10\,h^{-1}\,\mathrm{Mpc}, and reduce the variance of the estimator about fivefold.

DOI · arXiv:1607.01026

Large-scale structureSimulations

A 7h-1Mpc-thick slice through Horizon Run 4 at z = 0, with two successive zooms onto a galaxy cluster.
A 7 h−1 Mpc7\,h^{-1}\,\mathrm{Mpc}-thick slice through Horizon Run 4 at z=0z = 0, with two successive zooms onto a galaxy cluster.

2015 · J. Korean Astron. Soc. 48 (2015) 213

Horizon Run 4 simulation: coupled evolution of galaxies and large-scale structures of the Universe

Kim, J., Park, C., L'Huillier, B., Hong, S. E.

Horizon Run 4 follows 630036300^3 particles in a 3150 h−1 Mpc3150\,h^{-1}\,\mathrm{Mpc} box, with halo merger trees down to 2.7×1011 h−1 M⊙2.7\times10^{11}\,h^{-1}\,M_\odot. The halo mass function departs from universality and evolves with redshift, and the BAO peak in mock galaxy correlations broadens and shifts. The data are public.

DOI · arXiv:1508.05107

Key papersLarge-scale structureDark matter haloesSimulations

← All papers