Denoising Diffusion Probabilistic Model for realistic and fast generated \textit{Euclid}-like data for weak lensing analysis
arXiv:2504.07183 · doi:10.3847/1538-4357/adcec4
Abstract
Understanding and mitigating measurement systematics in weak lensing (WL) analysis requires large datasets of realistic galaxies with diverse morphologies and colors. Missions like Euclid, the Nancy Roman Space Telescope, and Vera C. Rubin Observatory's Legacy Survey of Space and Time will provide unprecedented statistical power and control over systematic uncertainties. Achieving the stringent shear measurement requirement of demands analyzing galaxies. Accurately modeling galaxy morphology is crucial, as it is shaped by complex astrophysical processes that are not yet fully understood. Subtle deviations in shape and structural parameters can introduce biases in shear calibration. The interplay between bulges, disks, star formation, and mergers contributes to morphological diversity, requiring simulations that faithfully reproduce these features to avoid systematics in shear measurements. Generating such a large and realistic dataset efficiently is feasible using advanced generative models like denoising diffusion probabilistic models (DDPMs). In this work, we extend Hubble Space Telescope (HST) data across Euclid's broad optical band using CANDELS and develop a generative AI tool to produce realistic Euclid-like galaxies while preserving morphological details. We validate our tool through visual inspection and quantitative analysis of galaxy parameters, demonstrating its capability to simulate realistic Euclid galaxy images, which will address WL challenges and enhance calibration for current and future cosmological missions.
9 pages, 5 figures, accepted for publication in ApJ
References in corpus (18)
- Astropy: A Community Python Package for Astronomy
- Binary companions of evolved stars in APOGEE DR14: Search method and catalog of ~5,000 companions
- LSST: from Science Drivers to Reference Design and Anticipated Data Products
- CANDELS: The Cosmic Assembly Near-infrared Deep Extragalactic Legacy Survey
- CANDELS: The Cosmic Assembly Near-infrared Deep Extragalactic Legacy Survey - The Hubble Space Telescope Observations, Imaging Data Products and Mosaics
- Diffusion Models in Vision: A Survey
- CANDELS Multi-wavelength Catalogs: Source Detection and Photometry in the GOODS-South Field
- Weak lensing shear calibration with simulations of the HSC survey
- Calibration of weak-lensing shear in the Kilo-Degree Survey
- DES Y3 results: Blending shear and redshift biases in image simulations
- Defining a weak lensing experiment in space
- Towards emulating cosmic shear data: Revisiting the calibration of the shear measurements for the Kilo-Degree Survey
- Deep Generative Models for Galaxy Image Simulations
- AstroVaDEr: Astronomical Variational Deep Embedder for Unsupervised Morphological Classification of Galaxies and Synthetic Image Generation
- Photometric Redshift Calibration Requirements for WFIRST Weak Lensing Cosmology: Predictions from CANDELS
- Euclid preparation: XIII. Forecasts for galaxy morphology with the Euclid Survey using Deep Generative Models
- Deblending Galaxies with Generative Adversarial Networks
- Euclid preparation. LXVII. Deep learning true galaxy morphologies for weak lensing shear bias calibration