Inferring subhalo effective density slopes from strong lensing observations with neural likelihood-ratio estimation

Zhang, Gemma; Mishra-Sharma, Siddharth; Dvorkin, Cora

doi:10.1093/mnras/stac3014

Cited by 5 publications

(1 citation statement)

References 89 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For such large datasets, the traditional forward modeling techniques would be unfeasible due to either computational or human time restrictions. Finally, MLbased inference provides a direct and fast way to constrain quantities of interest that can be otherwise too cumbersome to infer from the data through a traditional approach, e.g., detecting individual or populations of dark subhalos or constraining the subhalo-mass function (Brehmer et al, 2019;Coogan et al, 2020Ostdiek et al, 2020Ostdiek et al, , 2022Vernardos et al, 2020;Zhang et al, 2022;Anau Montel et al, 2022). Note that this particular science application is discussed in detail in Vegetti et al (in preparation).…”

Section: Machine-learning Based Parameter Extractionmentioning

confidence: 99%

Strong Lensing by Galaxies

Shajib¹,

Vernardos²,

Collett³

et al. 2022

Preprint

View full text Add to dashboard Cite

Strong lensing at the galaxy scale is useful for numerous applications in Astrophysics and Cosmology. Some of the principal applications are studying the mass structure, formation, and evolution of elliptical galaxies, constraining the stellar initial mass function, and measuring cosmological parameters. Since the first discovery of a galaxy-scale strong lens in the eighties, this field has come a long way regarding data quality and techniques to model the data. In this review article, we describe the most common methodologies to model lensing observables of galaxy-scale strong lenses, especially the imaging data, as it is the most available and informative source of lensing observables. We review the main results from the literature in astrophysical and cosmological applications of galaxy-scale strong lenses. We also discuss the current limitations of the data and methodologies and provide a future outlook of the expected development and improvements in both aspects in the near future.

show abstract

Section: Machine-learning Based Parameter Extractionmentioning

confidence: 99%

Strong Lensing by Galaxies

Shajib¹,

Vernardos²,

Collett³

et al. 2022

Preprint

View full text Add to dashboard Cite

show abstract

The effect of the perturber population on subhalo measurements in strong gravitational lenses

Coogan,

Anau Montel,

Karchev

et al. 2023

Monthly Notices of the Royal Astronomical Society

View full text Add to dashboard Cite

Analyses of extended arcs in strong gravitational lensing images to date have constrained the properties of dark matter by measuring the parameters of one or two individual subhalos. However, since such analyses are reliant on likelihood-based methods like Markov-chain Monte Carlo or nested sampling, they require various compromises to the realism of lensing models for the sake of computational tractability, such as ignoring the numerous other subhalos and line-of-sight halos in the system, assuming a particular form for the source model and requiring the noise to have a known likelihood function. Here we show that a simulation-based inference method called truncated marginal neural ratio estimation (TMNRE) makes it possible to relax these requirements by training neural networks to directly compute marginal posteriors for subhalo parameters from lensing images. By performing a set of inference tasks on mock data, we verify the accuracy of TMNRE and show it can compute posteriors for subhalo parameters marginalized over populations of hundreds of substructures, as well as lens and source uncertainties.We also find the multilayer perceptron (MLP) Mixer network works far better for such tasks than the convolutional architectures explored in other lensing analyses. Furthermore, we show that since TMNRE learns a posterior function it enables direct statistical checks that would be extremely expensive with likelihood-based methods. Our results show that TMNRE is well-suited for analyzing complex lensing data, and that the full subhalo and line-of-sight halo population must be included when measuring the properties of individual dark matter substructures with this technique.

show abstract

Anisotropic strong lensing as a probe of dark matter self-interactions

Dhanasingham,

Cyr-Racine,

Mace

et al. 2023

Monthly Notices of the Royal Astronomical Society

View full text Add to dashboard Cite

Galaxy-scale strongly lensed systems have been shown to provide a unique technique for exploring the underlying physics of dark matter at sub-galactic scales. In the past, much attention was given to detecting and studying individual haloes in a strong lens system. In addition to the subhaloes, line-of-sight haloes contribute significantly to the small perturbations in lensed images. In prior work, we demonstrated that these line-of-sight haloes imprint a distinctive anisotropic signature and hence give rise to a detectable non-zero parity-even quadrupole moment in the effective convergence field’s two-point correlation function. In this study, we show that these line-of-sight haloes also produce a non-zero curl component of the effective deflection field with a parity-odd quadrupole moment of the two-point function. These multipole moments have the ability to statistically separate line-of-sight haloes from dark matter substructure. In this paper, we examine how these multipole moments evolve in the presence of warm dark matter and self-interacting dark matter in terms of central density evolution and dark matter halo abundance. Importantly, we show that these different multipole moments display exquisite sensitivity to both the amplitude and the velocity dependence of the dark matter self-interaction cross-section. Our approach opens the door for strong lensing observations to probe dark matter self-interaction over a broad range of relative velocities.

show abstract

Inferring subhalo effective density slopes from strong lensing observations with neural likelihood-ratio estimation

Cited by 5 publications

References 89 publications

Strong Lensing by Galaxies

Strong Lensing by Galaxies

The effect of the perturber population on subhalo measurements in strong gravitational lenses

Anisotropic strong lensing as a probe of dark matter self-interactions

Contact Info

Product

Resources

About