Denoising weak lensing mass maps with diffusion model: systematic comparison with generative adversarial network
arXiv:2505.00345 · doi:10.1093/pasj/psag052
Abstract
Removing the shape noise from the observed weak lensing field, i.e., denoising, enhances the potential of WL by accessing information at small scales where the shape noise dominates without denoising. We utilise two machine learning (ML) models for denosing: generative adversarial network (GAN) and diffusion model (DM). We evaluate the performance of denosing with GAN and DM utilising the large suite of mock WL observations, which serve as the training and test data sets. We apply denoising to 1,000 noisy mass maps with GAN and DM models trained with 39,000 mock observations. Both models can fairly well reproduce the true convergence map on large scales. Then, we measure cosmological statistics: power spectrum, bispectrum, one-point probability distribution function, peak and minima counts, and scattering transform coefficients. We find that DM outperforms GAN in almost all considered statistics and recovers the correct statistics down to small scales. For example, the angular power spectrum can be recovered with DM up to multipoles while the noise power spectrum dominates from . We also conduct stress tests on the trained model; denoising the maps with different characteristics, e.g., different source redshifts, from the training data. The performance degrades at small scales, but the statistics can still be recovered at large scales. Though the training of DM is more computationally demanding compared with GAN, there are several advantages: numerically stable training, higher performance in the reconstruction of cosmological statistics, and sampling multiple realisations once the model is trained. It has been known that DM can generate higher-quality images in real-world problems than GAN, the superiority has been confirmed as well in the WL denoising problem.
PASJ in press, 18 pages, 20 figures, 5 tables
References in corpus (33)
- The Hyper Suprime-Cam SSP Survey: Overview and Survey Design
- The Dark Energy Survey: more than dark energy - an overview
- KiDS-450: Cosmological parameter constraints from tomographic weak gravitational lensing
- KiDS-1000 Cosmology: Multi-probe weak gravitational lensing and spectroscopic galaxy clustering constraints
- Euclid. I. Overview of the Euclid mission
- Cosmology with cosmic shear observations: a review
- Weak lensing for precision cosmology
- Cosmological parameters from lensing power spectrum and bispectrum tomography
- Evidence for the accelerated expansion of the Universe from weak lensing tomography with COSMOS
- Ray-tracing through the Millennium Simulation: Born corrections and lens-lens coupling in cosmic shear and galaxy-galaxy lensing
- Hyper Suprime-Cam Year 3 Results: Cosmology from Cosmic Shear Power Spectra
- Fast Estimators for Redshift-Space Clustering
- Searching for massive clusters in weak lensing surveys
- A new approach to observational cosmology using the scattering transform
- KiDS-Legacy: Cosmological constraints from cosmic shear with the complete Kilo-Degree Survey
- Cosmological Information in Weak Lensing Peaks
- Weak Lensing Mass Reconstruction using Wavelets
- Deep learning dark matter map reconstructions from DES SV weak lensing data
- A large sample of shear selected clusters from the Hyper Suprime-Cam Subaru Strategic Program S16A wide field mass maps
- Probing cosmology with weak lensing selected clusters I: Halo approach and all-sky simulations
- Weak Lensing Minima and Peaks: Cosmological Constraints and the Impact of Baryons
- TNG: Effect of Baryonic Processes on Weak Lensing with IllustrisTNG Simulations
- Denoising Weak Lensing Mass Maps with Deep Learning
- Probabilistic Mass Mapping with Neural Score Estimation
- Cosmology from weak lensing peaks and minima with Subaru Hyper Suprime-Cam survey first-year data
- Hundreds of weak lensing shear-selected clusters from the Hyper Suprime-Cam Subaru Strategic Program S19A data
- Deep Learning for Line Intensity Mapping Observations: Information Extraction from Noisy Maps
- A comparative study of cosmological constraints from weak lensing using Convolutional Neural Networks
- Wavelet Helmholtz decomposition for weak lensing mass map reconstruction
- Statistical properties of filaments in weak gravitational lensing
- A tomographic spherical mass map emulator of the KiDS-1000 survey using conditional generative adversarial networks
- Generative modelling for mass-mapping with fast uncertainty quantification
- Neural style transfer of weak lensing mass maps