Search arXiv⌕ Search
Home/Archive/arXiv

arXiv · 2610.37596

A self-supervised machine learning approach to identify instrumental systematics in 21cm Cosmology

Ajith Sampath
View original source ↗Read source document ↗Export metadata ↓

Abstract

The primordial sound waves in the baryon-photon plasma are imprinted on the large-scale galaxy distribution as Baryon Acoustic Oscillations (BAO), serving as a cosmological standard ruler to constrain dark energy and the Hubble parameter. The upcoming 21cm intensity-mapping arrays are designed to map the 21cm emission across large volumes and measure the BAO scale. However, radio interferometers suffer from instrumental systematics that can hinder detection of the cosmological 21cm signal by leaking bright foregrounds into the signal-dominated modes. This work aims to demonstrate that a self-supervised model can predict the instrument systematic state (here, the amplitude of a spectral ripple in the beam width) in interferometric visibilities and potentially disentangle the instrument from the sky without a forward model. Based on the fact that an instrument stays the same as the sky rotates in a drift-scan observation, and that the same sky is observed under different instrument states, we train a convolutional autoencoder to compress the information in the simulated visibilities and separate this information into instrument-only and sky-only latent vectors. The result shows excellent prediction of the unknown instrument systematic for 15 held-out instrument states with $R^2 = 0.9975$ with a linear probe, against $R^2 = -3.55$ for a random-encoder baseline. The instrument-latent carries no detectable sky signal, as measured with a total-power sky proxy, for any held-out test data. The sky-latent vector and instrument-latent vector show partial separation but not complete disentanglement, identifying sky-latent purity as the primary open problem. The results show the ability of the model to recognize the systematic and show a promising path towards an ML-based foreground cleaning pipeline. Full instrument-sky disentanglement constitutes the natural next step.

Explore related subjects

astro-ph.CO

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Ajith Sampath. 2026-10-27. A self-supervised machine learning approach to identify instrumental systematics in 21cm Cosmology. https://arxiv.org/abs/2610.37596

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related papers

arXiv2026

Increasing the Precision of Surrogate Models for Weak Lensing Mass Maps with Flow Matching

Weak gravitational lensing maps compactly encode the evolution of cosmic large-scale structure and are a key tool for cosmological analyses. Performing inference directly at the map level allows flexible choices of statistics and can increase constraining power. Conventional methods rely solely on N-body simulations and are computationally expensive. Generative machine-learning emulators can accelerate map-level theory prediction. However, existing GAN-based map-level surrogates still have limited statistical fidelity. They can produce over-smoothed maps, may fail to capture the full distribution of generated map sets and can be difficult to train. Continuous normalizing flows trained with flow matching have recently emerged as a powerful class of generative models. We present a residual label-conditional flow matching generative network that conditions explicitly on the matter density $Ω_m$ and clustering amplitude $σ_8$ for a fixed source redshift distribution $n(z)$. The model learns a continuous probability flow in a residual space from label-specific noise distributions to convergence maps. We evaluate it using pixel and peak statistics, the power spectrum, bispectrum, power-spectrum correlation matrices, and other validation metrics. Compared with the previous GAN benchmark, the proposed method improves the typical fidelity of generated maps from below 10% and below 20% to below 1% and below 5% for basic and higher-order statistics, respectively. The agreement at the level of map distributions is also very good: maps generated from random noise match well the distribution of maps generated with N-body simulations from random initial conditions. This work brings us closer to a practical mass-map emulator that captures the cosmological signal while supporting multiple forms of data analysis.

astro-ph.CO↗
arXiv2026

Pushing the Primordial Frontier: Exact Linear Solutions in Multifield Inflation

We present exact analytic solutions for the linear dynamics of a two-field inflationary system in which the primordial curvature perturbation $ζ$ is coupled to an isocurvature perturbation $σ$ of entropy mass $μ$. The solutions are valid for arbitrary values of $μ$ and the dimensionless interaction strength $λ$. They therefore provide analytic control over the strong mixing regime in which $ζ$ interacts with light isocurvature fields, commonly associated with rapid-turn inflation. As a first application, we derive the amplitude of the primordial power spectrum in closed form, obtaining an expression that interpolates between the weakly mixed, strongly mixed, light-field, and heavy-field regimes. These results open the way to analytic studies of multifield observables beyond the power spectrum, including non-Gaussianity, particle production, and loop corrections.

astro-ph.CO↗
arXiv2026

EFT-Ramses: a code to simulate the effective field theory of dark energy

While the standard $Λ$CDM paradigm is in excellent agreement with most current cosmological observations, theoretical challenges surrounding the cosmological constant ($Λ$) have strongly motivated the exploration of dynamical dark energy (DE) and modified gravity (MG) models. Investigating the physical nature of the cosmic acceleration requires N-body simulations to probe the non-linear growth of cosmic structure and prepare for the high-precision data from Stage-IV surveys. In this paper, we present EFT-RAMSES, a comprehensive extension of the ECOSMOG cosmological simulation code designed to explore non-linear structure formation in DE and MG scenarios. We embed the effective field theory of dark energy (EFTofDE) framework into this new numerical pipeline, utilising the $α$-basis parameterisation to provide a versatile, model-agnostic, computational engine. By consolidating diverse scalar and vector-tensor theories---including the normal and self-accelerating Dvali-Gabadadze-Porrati (DGP) models, cubic Galileons (cubic scalar Galileon (csG), cubic vector Galileon (cvG) and generalised cubic covariant Galileon (GCCG)), and generic effective field theory (EFT) parameterisation---into a single "master" Vainshtein equation, this pipeline bypasses the need for model-specific solvers and easily specialises to any particular model. As validations, we perform high-resolution N-body simulations for the normal-branch DGP (nDGP), csG, GCCG, and EFT models, comparing the resulting matter power spectra against dependent and independent codes such as legacy ECOSMOG and HiCOLA, as well as linear theory, and find excellent agreement. EFT-RAMSES provides a robust and versatile computational tool for precision cosmological tests of DE and MG using upcoming cosmological surveys. The code is available for download from the GitHub EFT-RAMSES repository.

astro-ph.CO↗