Optimal, fast, and robust inference of reionization-era cosmology with the 21cmPIE-INN

Abstract

Modern machine learning will allow for simulation-based inference from reionization-era 21cm observations at the Square Kilometre Array. Our framework combines a convolutional summary network and a conditional invertible network through a physics-inspired latent representation. It allows for an efficient and extremely fast determination of the posteriors of astrophysical and cosmological parameters, jointly with well-calibrated and on average unbiased summaries. The sensitivity to non-Gaussian information makes our method a promising alternative to the established power spectra.

15+10 pages, 11 figures

References in corpus (30)

• Efficient Simulations of Early Structure Formation and Reionization
• Guided Image Generation with Conditional Invertible Neural Networks
• Real-time gravitational-wave science with neural posterior estimation
• 21cmFAST v3: A Python-integrated C code forgenerating 3D realizations of the cosmic 21cm signal
• Machine Learning and LHC Event Generation
• The DAWES review 10: The impact of deep learning for the analysis of galaxy surveys
• New Constraints on Warm Dark Matter from the Lyman-$\alpha$ Forest Power Spectrum
• Unveiling Dark Matter free-streaming at the smallest scales with high redshift Lyman-alpha forest
• Unveiling the nature of dark matter with high redshift 21 cm line experiments
• Simulation-Based Inference of Reionization Parameters From 3D Tomographic 21 cm Lightcone Images
• MadNIS -- Neural Multi-Channel Importance Sampling
• Cosmology with one galaxy?
• Two Invertible Networks for the Matrix Element Method
• ERGO-ML I: Inferring the assembly histories of IllustrisTNG galaxies from integral observable properties via invertible neural networks
• Exploring the likelihood of the 21-cm power spectrum with simulation-based inference
• Modern Machine Learning for LHC Physicists
• Machine Learning for Observational Cosmology
• Constraining the X-ray heating and reionization using 21-cm power spectra with Marginal Neural Ratio Estimation
• SKA Science Data Challenge 2: analysis and results
• Inferring Astrophysics and Dark Matter Properties from 21cm Tomography using Deep Learning
• Cubic-Spline Flows
• Quantifying Uncertainty in Deep Learning Approaches to Radio Galaxy Classification
• Machine Learning and Cosmology
• Cornering Extended Starobinsky Inflation with CMB and SKA
• Synergistic tests of inflation
• Noise-Net: Determining physical properties of HII regions reflecting observational uncertainties
• Galaxy Spectra neural Network (GaSNet). II. Using Deep Learning for Spectral Classification and Redshift Predictions
• Deep Learning 21cm Lightcones in 3D
• 3D ScatterNet: Inference from 21 cm Light-cones
• Application-driven Validation of Posteriors in Inverse Problems

Cited by in corpus (9)

• SKATR: A Self-Supervised Summary Transformer for SKA
• The 21cm-galaxy cross-correlation: Realistic forecast for 21cm signal detection and reionisation constraints
• From ANN to BNN: Inferring Reionization Parameters using Uncertainty-aware Emulators of 21-cm Summaries
• Observable Optimization for Precision Theory: Machine Learning Energy Correlators
• 21 cm Cosmology Sensitivity to Small-Scale Structure: Warm vs Neutrino-Interacting Dark Matter
• Sample Variance Denoising in Cylindrical 21-cm Power Spectra
• Large Language Models -- the Future of Fundamental Physics?
• Direct reconstruction of the Reionization history from 21cm 2D Power Spectra
• CosmoUiT: A Vision Transformer-UNet Hybrid for Fast and Accurate Emulation of 21-cm Maps from the Epoch of Reionization