We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:


Current browse context:


Change to browse by:

References & Citations


(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Astrophysics > Cosmology and Nongalactic Astrophysics

Title: Weak lensing scattering transform: dark energy and neutrino mass sensitivity

Abstract: As weak lensing surveys become deeper, they reveal more non-Gaussian aspects of the convergence field which require statistics beyond the power spectrum to extract. In Cheng et al. (2020), we showed that the scattering transform, a novel statistic borrowing concepts from convolutional neural networks, is a powerful tool to perform cosmological parameter estimation. Here, we extend this analysis to explore its sensitivity to dark energy and neutrino mass parameters with weak lensing surveys. We first use image synthesis to show visually that the scattering transform provides a better statistical vocabulary to characterize lensing mass maps compared to the power spectrum and bispectrum. We then show that it outperforms those two estimators in the low-noise regime. When the noise level increases and non-Gaussianity is diluted, though the constraint is not significantly tighter than that of the bispectrum, the scattering coefficients have much more Gaussian likelihood, which is essential for accurate cosmological inference. We argue that the scattering coefficients are preferred statistics considering both constraining power and likelihood properties.
Comments: 8 pages, 6 figures
Subjects: Cosmology and Nongalactic Astrophysics (astro-ph.CO); Instrumentation and Methods for Astrophysics (astro-ph.IM)
Cite as: arXiv:2103.09247 [astro-ph.CO]
  (or arXiv:2103.09247v1 [astro-ph.CO] for this version)

Submission history

From: Sihao Cheng [view email]
[v1] Tue, 16 Mar 2021 18:00:06 GMT (4186kb,D)
[v2] Thu, 8 Apr 2021 14:42:49 GMT (4184kb,D)
[v3] Thu, 22 Jul 2021 18:00:03 GMT (4186kb,D)

Link back to: arXiv, form interface, contact.