Deep learning approach for identification of H <scp>ii</scp> regions during reionization in 21-cm observations

Bianco, Michele; Giri, Sambit K.; Iliev, Ilian T.; Mellema, Garrelt

doi:10.1093/mnras/stab1518

Cited by 31 publications

(18 citation statements)

References 86 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…They can be related to several other quantities that have been used to study reionization. First of all, the Betti numbers 𝛽 𝑖 (𝛼) for 𝑖 ∈ {0, 1, 2} count the number of features alive as a function of 𝛼 (Elbers 2017;Elbers & van de Weygaert 2019;Kapahtia et al 2018Kapahtia et al , 2019Kapahtia et al , 2021Bianco et al 2021) and are therefore an 'integral' of the persistence diagrams. The commonly used Euler characteristic (Lee et al 2008;Friedrich et al 2011;Hong et al 2014) is an alternating sum of Betti numbers, 𝜒 = 𝛽 0 − 𝛽 1 + 𝛽 2 , and one of the Minkowski functionals (Gleser et al 2006;Yoshiura et al 2016;Kapahtia et al 2018;Bag et al 2018;Chen et al 2019).…”

Section: Discussionmentioning

confidence: 99%

“…Besides our earlier work (Elbers 2017, Paper I), Betti numbers have been used in the context of reionization by Kapahtia et al (2018Kapahtia et al ( , 2019Kapahtia et al ( , 2021; and Bianco et al (2021). Among these, the work of is most closely related to our own, while Kapahtia et al (2018Kapahtia et al ( , 2019Kapahtia et al ( , 2021 analyse two-dimensional temperature maps.…”

Section: Persistent Homologymentioning

confidence: 99%

See 1 more Smart Citation

Persistent topology of the reionisation bubble network. I: Formalism & Phenomenology

Elbers

Weygaert

2019

Monthly Notices of the Royal Astronomical Society

View full text Add to dashboard Cite

We present a new formalism for studying the topology of H ii regions during the Epoch of Reionisation, based on persistent homology theory. With persistent homology, it is possible to follow the evolution of topological features over time. We introduce the notion of a persistence field as a statistical summary of persistence data and we show how these fields can be used to identify different stages of reionisation. We identify two new stages common to all bubble ionisation scenarios. Following an initial pre-overlap and subsequent overlap stage, the topology is first dominated by neutral filaments (filament stage) and then by enclosed patches of neutral hydrogen undergoing outsidein ionisation (patch stage). We study how these stages are affected by the degree of galaxy clustering. We also show how persistence fields can be used to study other properties of the ionisation topology, such as the bubble size distribution and the fractal-like topology of the largest ionised region.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Persistent Homologymentioning

confidence: 99%

Persistent topology of the reionisation bubble network. I: Formalism & Phenomenology

Elbers

Weygaert

2019

Monthly Notices of the Royal Astronomical Society

View full text Add to dashboard Cite

show abstract

“…These include the one-point statistics of the brightness temperature (Watkinson & Pritchard 2014;Shimabukuro et al 2015;Kubota et al 2016;Banet et al 2021;Gorce et al 2021), the morphological and/or topological features of the 21-cm signal (e.g. Yoshiura et al 2017;Bag et al 2019;Chen et al 2019;Elbers & van de Weygaert 2019;Kapahtia et al 2019;Gazagnes et al 2021;Kapahtia et al 2021) and the distribution of the sizes of ionised regions (Kakiichi et al 2017;Giri et al 2018aGiri et al ,b, 2019bBianco et al 2021). Alternatively, convolutional neural networks (CNNs) have been considered which are trained to be able to extract 2D or 3D features from images of the cosmic 21-cm signal and used to extract astrophysical information from mock observations (e.g.…”

Section: Introductionmentioning

confidence: 99%

Exploring the cosmic 21-cm signal from the Epoch of Reionisation using the Wavelet Scattering Transform

Greig,

Ting,

Kaurov

2022

Preprint

View full text Add to dashboard Cite

Detecting the cosmic 21-cm signal during the Epoch of Reionisation and Cosmic Dawn will reveal insights into the properties of the first galaxies and advance cosmological parameter estimation. Until recently, the primary focus for astrophysical parameter inference from the 21-cm signal centred on the power spectrum (PS). However, the cosmic 21-cm signal is highly non-Gaussian rendering the PS sub-optimal for characterising the cosmic signal. In this work, we introduce a new technique to analyse the non-Gaussian information in images of the 21-cm signal called the Wavelet Scattering Transform (WST). This approach closely mirrors that of convolutional neural networks with the added advantage of not requiring tuning or training of a neural network. Instead, it compresses the 2D spatial information into a set of coefficients making it easier to interpret while also providing a robust statistical description of the non-Gaussian information contained in the cosmic 21-cm signal. First, we explore the application of the WST to mock 21-cm images to gain valuable physical insights by comparing to the known behaviour from the 21-cm PS. Then we quantitatively explore the WST applied to the 21-cm signal by extracting astrophysical parameter constraints using Fisher Matrices from a realistic 1000 hr mock observation with the Square Kilometre Array. We find that: (i) the WST applied only to 2D images can outperform the 3D spherically averaged 21-cm PS, (ii) the excision of foreground contaminated modes can degrade the constraining power by a factor of ∼ 1.5 − 2 with the WST and (iii) higher cadences between the 21-cm images can further improve the constraining power.

show abstract

“…It is also sensitive to the non-Gaussian information in the signal. While instrumental limitations, starting with the increasing thermal noise at higher angular resolution (e.g., Mellema et al 2013), will affect the process, methods for obtaining the bubble size distribution from tomographic observations have been explored (Giri et al 2018;Bianco et al 2021). Semi-numerical, or even numerical simulations, could then be used to perform parameter inference studies based on this quantity.…”

Section: Introductionmentioning

confidence: 99%

A bubble size distribution model for the Epoch of Reionization

Doussot

Semelin

2022

A&A

View full text Add to dashboard Cite

Aims. The bubble size distribution is a summary statistics that can be computed from the observed 21-cm signal from the Epoch of Reionization. As it depends only on the ionization field and is not limited to gaussian information, it is an interesting probe, complementary to the power spectrum of the full 21-cm signal. Devising a flexible and reliable theoretical model for the bubble size distribution paves the way for using it for astrophysical parameters inference. Methods. The proposed model is built from the excursion set theory and a functional relation between the bubble volume and the collapsed mass in the bubble. Unlike previous models it accommodates any functional relation or distributions. Using parameterized relations allows us to test the predictive power of the model by performing a minimization best-fit to the bubble size distribution obtained from a high resolution, fully coupled radiative hydrodynamics simulations, HIRRAH-21. Results. Our model is able to provide a better fit to the numerical bubble size distribution at ionization fraction of x HII ∼ 1% and 3% than other existing models. Moreover, the bubble volume to collapsed mass relation corresponding to the best-fit parameters, which is not an observable, is compared to numerical simulation data. A good match is obtained, confirming the possibility to infer this relation from an observed bubble size distribution using our model. Finally we present a simple algorithm that empirically implements the process of percolation. We show that it extends the usability of our bubble size distribution model up to x HII ∼ 30%.

show abstract

Deep learning approach for identification of H ii regions during reionization in 21-cm observations

Cited by 31 publications

References 86 publications

Persistent topology of the reionisation bubble network. I: Formalism & Phenomenology

Persistent topology of the reionisation bubble network. I: Formalism & Phenomenology

Exploring the cosmic 21-cm signal from the Epoch of Reionisation using the Wavelet Scattering Transform

A bubble size distribution model for the Epoch of Reionization

Contact Info

Product

Resources

About