CoLFI: Cosmological Likelihood-free Inference with Neural Density Estimators

doi:10.3847/1538-4365/ace113

	CoLFI: Cosmological Likelihood-free Inference with Neural Density Estimators
	Wang, Guo-Jian 1,2,3; Cheng, Cheng1,2,4 ; Ma, Yin-Zhe 1,2,3,5; Xia, Jun-Qing 6; Abebe, Amare 3,7; Beesham, Aroonkumar 3,8,9
	2023-09-01
Source Publication	ASTROPHYSICAL JOURNAL SUPPLEMENT SERIES
ISSN	0067-0049
Volume	268 Issue:1 Pages:7
Contribution Rank	4
Abstract	In previous works, we proposed to estimate cosmological parameters with an artificial neural network (ANN) and a mixture density network (MDN). In this work, we propose an improved method called a mixture neural network (MNN) to achieve parameter estimation by combining ANN and MDN, which can overcome shortcomings of the ANN and MDN methods. Besides, we propose sampling parameters in a hyperellipsoid for the generation of the training set, which makes the parameter estimation more efficient. A high-fidelity posterior distribution can be obtained using O(10(2)) forward simulation samples. In addition, we develop a code named CoLFI for parameter estimation, which incorporates the advantages of MNN, ANN, and MDN, and is suitable for any parameter estimation of complicated models in a wide range of scientific fields. CoLFI provides a more efficient way for parameter estimation, especially for cases where the likelihood function is intractable or cosmological models are complex and resource-consuming. It can learn the conditional probability density p(?\|d(0)) using samples generated by models, and the posterior distribution p(?\|d(0)) can be obtained for a given observational data d(0). We tested the MNN using power spectra of the cosmic microwave background and Type Ia supernovae and obtained almost the same result as the Markov Chain Monte Carlo method. The numerical difference only exists at the level of 0 (10(-2)s ). The method can be extended to higher-dimensional data.
Corresponding Author	Ma, Yin-Zhe(ma@ukzn.ac.za)
DOI	10.3847/1538-4365/ace113
Indexed By	SCI
Language	英语
WOS Keyword	APPROXIMATE BAYESIAN COMPUTATION ; REIONIZATION PARAMETERS ; CONSTRAINTS ; IA
Funding Project	National Research Foundation[150580] ; National Research Foundation[120385] ; National Research Foundation[120378] ; National Science Foundation of China[U1931202]
WOS Research Area	Astronomy & Astrophysics
WOS Subject	Astronomy & Astrophysics
WOS ID	WOS:001052840400001
Publisher	IOP Publishing Ltd
Funding Organization	National Research Foundation ; National Science Foundation of China
Citation statistics	Cited Times:5[WOS] [WOS Record] [Related Records in WOS]
Document Type	期刊论文
Identifier	http://ir.xao.ac.cn/handle/45760611-7/5425
Collection	射电天文研究室_理论天体物理研究团组
Corresponding Author	Ma, Yin-Zhe
Affiliation	1.Univ KwaZulu Natal, Sch Chem & Phys, Westville Campus Private Bag X54001, ZA-4000 Durban, South Africa 2.Univ KwaZulu Natal, NAOC UKZN Comp Astrophys Ctr NUCAC, ZA-4000 Durban, South Africa 3.Natl Inst Theoret & Comp Sci NITheCS, Johannesburg, South Africa 4.Chinese Acad Sci, Xinjiang Astron Observ, Urumqi 830011, Peoples R China 5.Stellenbosch Univ, Dept Phys, ZA-7602 Matieland, South Africa 6.Beijing Normal Univ, Dept Astron, Beijing 100875, Peoples R China 7.North West Univ, Ctr Space Res, ZA-2520 Potchefstroom, South Africa 8.Univ Zululand, Dept Math Sci, Private Bag X1001, ZA-3886 Kwa Dlangezwa, South Africa 9.Mangosuthu Univ Technol, Fac Nat Sci, POB 12363, ZA-4052 Jacobs, South Africa
Recommended Citation GB/T 7714	Wang, Guo-Jian,Cheng, Cheng,Ma, Yin-Zhe,et al. CoLFI: Cosmological Likelihood-free Inference with Neural Density Estimators[J]. ASTROPHYSICAL JOURNAL SUPPLEMENT SERIES,2023,268(1):7.
APA	Wang, Guo-Jian,Cheng, Cheng,Ma, Yin-Zhe,Xia, Jun-Qing,Abebe, Amare,&Beesham, Aroonkumar.(2023).CoLFI: Cosmological Likelihood-free Inference with Neural Density Estimators.ASTROPHYSICAL JOURNAL SUPPLEMENT SERIES,268(1),7.
MLA	Wang, Guo-Jian,et al."CoLFI: Cosmological Likelihood-free Inference with Neural Density Estimators".ASTROPHYSICAL JOURNAL SUPPLEMENT SERIES 268.1(2023):7.

Files in This Item:
File Name/Size	DocType	Version	Access	License
Wang-2023-CoLFI_ Cos（3105KB）	期刊论文	出版稿	开放获取	CC BY-NC-SA	View Application Full Text

File name:	Wang-2023-CoLFI_ Cosmological Likelihood-free.pdf
Format:	Adobe PDF

If you have any objections to this item, please fill out the form below and the administrator will contact you as soon as possible.
Content:
Email：	*
Institution:
Verification Code:	Refresh

Any comments and suggestions are welcomed.
Title:	*
Content:
Email：	*
Verification Code:	Refresh

射电天文研究室知识库