Ly$\alpha$NNA: A Deep Learning Field-level Inference Machine for the Lyman-$\alpha$ Forest

Abstract

The inference of astrophysical and cosmological properties from the Lyman-$\alpha$ forest conventionally relies on summary statistics of the transmission field that carry useful but limited information. We present a deep learning framework for inference from the Lyman-$\alpha$ forest at field-level. This framework consists of a 1D residual convolutional neural network (ResNet) that extracts spectral features and performs regression on thermal parameters of the IGM that characterize the power-law temperature-density relation. We train this supervised machinery using a large set of mock absorption spectra from Nyx hydrodynamic simulations at $z=2.2$ with a range of thermal parameter combinations (labels). We employ Bayesian optimization to find an optimal set of hyperparameters for our network, and then employ a committee of 20 neural networks for increased statistical robustness of the network inference. In addition to the parameter point predictions, our machine also provides a self-consistent estimate of their covariance matrix with which we construct a pipeline for inferring the posterior distribution of the parameters. We compare the results of our framework with the traditional summary (PDF and power spectrum of transmission) based approach in terms of the area of the 68% credibility regions as our figure of merit (FoM). In our study of the information content of perfect (noise- and systematics-free) Ly$\alpha$ forest spectral data-sets, we find a significant tightening of the posterior constraints -- factors of 10.92 and 3.30 in FoM over power spectrum only and jointly with PDF, respectively -- that is the consequence of recovering the relevant parts of information that are not carried by the classical summary statistics.

15 pages, 16 figures, to be published in Astronomy and Astrophysics. This update contains the peer-reviewd, accepted version of the paper

References in corpus (15)

• Cosmological parameters from combining the Lyman-alpha forest with CMB, galaxy clustering and SN constraints
• Baryon Acoustic Oscillations in the Ly-α forest of BOSS quasars
• Model-independent evidence in favor of an end to reionization by z~6
• Evidence of patchy hydrogen reionization from an extreme Ly$\alpha$ trough below redshift six
• The Completed SDSS-IV extended Baryon Oscillation Spectroscopic Survey: Baryon acoustic oscillations with Lyman-$\alpha$ forests
• Possible evidence for an inverted temperature-density relation in the intergalactic medium from the flux distribution of the Lyman-alpha forest
• The thermal history of the intergalactic medium down to redshift z=1.5: a new curvature measurement
• IGM Constraints from the SDSS-III/BOSS DR9 Ly-alpha Forest Flux Probability Distribution Function
• Simulating intergalactic gas for DESI-like small scale Lymanα forest observations
• Machine Learning for Observational Cosmology
• Thermal state of the intergalactic medium at $z\sim 2-4$
• Modeling the Lyman-$\alpha$ forest with Eulerian and SPH hydrodynamical methods
• A Novel Statistical Method for Measuring the Temperature-Density Relation in the IGM Using the $b$-$N_{\text{HI}}$ Distribution of Absorbers in the Ly$\alpha$ Forest
• Improving IGM temperature constraints using wavelet analysis on high-redshift quasars
• Measuring the thermal and ionization state of the low-$z$ IGM using likelihood free inference

Cited by in corpus (3)

• Deep learning-driven likelihood-free parameter inference for 21-cm forest observations
• Towards characterizing dark matter subhalo perturbations in stellar streams with graph neural networks
• Parameter estimation from Ly$\alpha$ forest in Fourier space using Information Maximising Neural Network