Constraining the Reionization History using Bayesian Normalizing Flows

@article{Hortua2020ConstrainingTR,
  title={Constraining the Reionization History using Bayesian Normalizing Flows},
  author={H'ector J. Hort'ua and Luigi Malag{\`o} and Riccardo Volpi},
  journal={Machine Learning: Science and Technology},
  year={2020},
  volume={1}
}
Upcoming experiments such as Hydrogen Epoch of Reionization Array(HERA) and the Square Kilometre Array (SKA) are intended to measure the 21 cm signal over a wide range of redshifts, representing an incredible opportunity in advancing our understanding about the nature of cosmic reionization. At the same time these kind of experiments will present new challenges in processing the extensive amount of data generated, calling for the development of automated methods capable of precisely estimating… 

Deep-Learning Study of the 21-cm Differential Brightness Temperature During the Epoch of Reionization

We propose a deep learning analysis technique with a convolutional neural network (CNN) to predict the evolutionary track of the Epoch of Reionization (EoR) from the 21-cm differential brightness

The GCE in a New Light: Disentangling the γ-ray Sky with Bayesian Graph Convolutional Neural Networks

Bayesian neural networks have the potential to resolve the debate regarding the Galactic Center excess, and when applied to the Fermi photon-count map, the NN identifies a smooth GCE in the data, suggestive of the presence of DM.

Flow-based likelihoods for non-Gaussian inference

This work investigates the use of data-driven likelihoods to bypass a key assumption made in many scientific analyses, and suggests using the optimization targets of flow-based generative models, a class of models that can capture complex distributions by transforming a simple base distribution through layers of nonlinearities.

QuasarNet: A new research platform for the data-driven investigation of black holes

QuasarNet demonstrates the power of ML, in analyzing and exploring large datasets, and also offers a unique opportunity to interrogate the theoretical assumptions that underpin accretion and feedback models.

Constraining cosmological parameters from N-body simulations with Variational Bayesian Neural Networks

This work implements multiplicative normalizing (MNFs), a family of approximate posteriors for the parameters of BNNs with the purpose of enhancing the flexibility of the variational posterior distribution, to extract Ω m, h, and σ 8 from the QUIJOTE simulations.

Amortized Bayesian Inference for Supernovae in the Era of the Vera Rubin Observatory Using Normalizing Flows

This work presents the first simulation-based inference method using normalizingows, trained to rapidly infer the parameters of toy supernovae model in multivariate, Rubin-like datastreams, requiring only one-ten-thousandth of the CPU hours during test time.

Machine Learning and Cosmology

Current and ongoing developments relating to the application of machine learning within cosmology and a set of recommendations aimed at maximizing the scientific impact of these burgeoning tools over the coming decade through both technical development as well as the fostering of emerging communities are summarized.

Understanding the Impact of Semi-numeric Reionization Models when Using CNNs

It is shown that neural networks trained on input data from only one model produce poor predictions on data from the other model, and this finding has important implications for future analyzes on observation data, and encourages the use of multiple models to produce images that capture the full complexity of the EoR.

Simulation-based Inference of Reionization Parameters from 3D Tomographic 21 cm Light-cone Images

This work adopts a trained 3D convolutional neural network to compress the 3D image data into informative summaries (DELFI-3D CNN) and shows that this method recovers accurate posterior distributions for the reionization parameters.

References

SHOWING 1-10 OF 41 REFERENCES

Simultaneously constraining the astrophysics of reionisation and the epoch of heating with 21CMMC

Abstract We extend our MCMC sampler of 3D EoR simulations, 21CMMC, to perform parameter estimation directly on light-cones of the cosmic 21cm signal. This brings theoretical analysis one step closer

Constraining the astrophysics and cosmology from 21 cm tomography using deep learning with the SKA

Future Square Kilometre Array (SKA) surveys are expected to generate huge data sets of 21 cm maps on cosmological scales from the Epoch of Reionization. We assess the viability of exploiting

Parameters Estimation for the Cosmic Microwave Background with Bayesian Neural Networks

Flipout outperforms all other methods regardless of the architecture used, and provides tighter constraints for the cosmological parameters, and the correct calibration of these networks does not change the behavior for the aleatoric and epistemic uncertainties provided for BNNs when the size of the training dataset changes.

Identifying Reionization Sources from 21cm Maps using Convolutional Neural Networks

Active Galactic Nuclei (AGN) and star-forming galaxies are leading candidates for being the luminous sources that reionized our Universe. Next-generation 21cm surveys are promising to break

Parameters Estimation from the 21 cm signal using Variational Inference

Variational Inference, and in particular Bayesian Neural Networks, are employed as an alternative to MCMC in 21 cm observations to report credible estimations for cosmological and astrophysical parameters and assess the correlations among them.

Cosmological Parameter Estimation Using 21 cm Radiation from the Epoch of Reionization

A number of radio interferometers are currently being planned or constructed to observe 21 cm emission from reionization. Not only will such measurements provide a detailed view of that epoch, but,

21CMMC: an MCMC analysis tool enabling astrophysical parameter studies of the cosmic 21 cm signal

We introduce 21CMMC: a parallelized, Monte Carlo Markov Chain analysis tool, incorporating the epoch of reionization (EoR) seminumerical simulation 21CMFAST. 21CMMC estimates astrophysical parameter

Deep learning from 21-cm tomography of the Cosmic Dawn and Reionization

The 21-cm power spectrum (PS) has been shown to be a powerful discriminant of reionization and cosmic dawn astrophysical parameters. However, the 21-cm tomographic signal is highly non-Gaussian.

21cmfast: a fast, seminumerical simulation of the high‐redshift 21‐cm signal

We introduce a powerful semi-numeric modeling tool, 21cmFAST, designed to efficiently simulate the cosmological 21-cm signal. Our code generates 3D realizations of evolved density, ionization,