# Non-Gaussian information from weak lensing data via deep learning

arXiv: https://arxiv.org/abs/1802.01212 · HTML: https://mlsanity.com/paper/1802.01212/

## Key facts

- Cited by 11 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 4 Feb 2018.

## Built on by (most cited first, 11 of 11)

- [Multifield Cosmology with Artificial Intelligence](https://mlsanity.com/paper/2109.09747/) (2021)
- [Wavelet Conditional Renormalization Group](https://mlsanity.com/paper/2207.04941/) (2022)
- [Multiscale Flow for Robust and Optimal Cosmological Analysis](https://mlsanity.com/paper/2306.04689/) (2023)
- [Scattering Spectra Models for Physics](https://mlsanity.com/paper/2306.17210/) (2023)
- [A simple equivariant machine learning method for dynamics based on scalars](https://mlsanity.com/paper/2110.03761/) (2021)
- [Machine Learning and Cosmology](https://mlsanity.com/paper/2203.08056/) (2022)
- [Conditionally Strongly Log-Concave Generative Models](https://mlsanity.com/paper/2306.00181/) (2023)
- [FAIR Universe Weak Lensing ML Uncertainty Challenge: Handling Uncertainties and Distribution Shifts for Precision Cosmology](https://mlsanity.com/paper/2604.14451/) (2026)
- [Learnable wavelet neural networks for cosmological inference](https://mlsanity.com/paper/2307.14362/) (2023)
- [Dark Energy Survey Year 3 results: Simulation-based wCDM inference from weak lensing and galaxy clustering maps with deep learning: Analysis design](https://mlsanity.com/paper/2511.04681/) (2025)
- [Generative Modeling via Kernelized Stochastic Interpolants](https://mlsanity.com/paper/2602.20070/) (2026)

## Abstract (arXiv)

Weak lensing maps contain information beyond two-point statistics on small scales. Much recent work has tried to extract this information through a range of different observables or via nonlinear transformations of the lensing field. Here we train and apply a 2D convolutional neural network to simulated noiseless lensing maps covering 96 different cosmological models over a range of Omegam,σ8. Using the area of the confidence contour in the Omegam,σ8 plane as a figure-of-merit, derived from simulated convergence maps smoothed on a scale of 1.0 arcmin, we show that the neural network yields ≈ 5 × tighter constraints than the power spectrum, and ≈ 4 × tighter than the lensing peaks. Such gains illustrate the extent to which weak lensing data encode cosmological information not accessible to the power spectrum or even other, non-Gaussian statistics such as lensing peaks.