Likelihood-free Model Selection in Cosmic Reionization with Three-dimensional Tomographic 21 cm Lightcone Images

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Binnie, T., Zhao, Xiaosheng, Pritchard, J. R., Mao, Yi
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866915147949277184
author Binnie, T.
Zhao, Xiaosheng
Pritchard, J. R.
Mao, Yi
author_facet Binnie, T.
Zhao, Xiaosheng
Pritchard, J. R.
Mao, Yi
contents We explore likelihood-free (aka simulation-based) Bayesian model selection to quantify model comparison analyses of reionisation scenarios. We iteratively train the 3D Convolutional Neural Network (CNN) on four toy EoR models based on 21cmFAST simulations with contrasting morphology to obtain summaries of the 21 cm lightcone. Within the pyDelfi framework, we replaced the Emcee sampler with MultiNest to integrate learnt posteriors and produce the Bayesian Evidence. We comfortably distinguish the model used to produce the mock data set in all cases. However, we struggle to produce accurate posterior distributions for outside-in reionisation models. After a variety of cross-checks and alternate analyses we discuss the flexibility of summarising models that differ from precisely the intended network training conditions as this should be more widely scrutinised before CNN can reliably analyse observed data.
format Preprint
id arxiv_https___arxiv_org_abs_2502_08152
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Likelihood-free Model Selection in Cosmic Reionization with Three-dimensional Tomographic 21 cm Lightcone Images
Binnie, T.
Zhao, Xiaosheng
Pritchard, J. R.
Mao, Yi
Instrumentation and Methods for Astrophysics
Cosmology and Nongalactic Astrophysics
We explore likelihood-free (aka simulation-based) Bayesian model selection to quantify model comparison analyses of reionisation scenarios. We iteratively train the 3D Convolutional Neural Network (CNN) on four toy EoR models based on 21cmFAST simulations with contrasting morphology to obtain summaries of the 21 cm lightcone. Within the pyDelfi framework, we replaced the Emcee sampler with MultiNest to integrate learnt posteriors and produce the Bayesian Evidence. We comfortably distinguish the model used to produce the mock data set in all cases. However, we struggle to produce accurate posterior distributions for outside-in reionisation models. After a variety of cross-checks and alternate analyses we discuss the flexibility of summarising models that differ from precisely the intended network training conditions as this should be more widely scrutinised before CNN can reliably analyse observed data.
title Likelihood-free Model Selection in Cosmic Reionization with Three-dimensional Tomographic 21 cm Lightcone Images
topic Instrumentation and Methods for Astrophysics
Cosmology and Nongalactic Astrophysics
url https://arxiv.org/abs/2502.08152