中国科学:物理学 力学 天文学(英文版)2024，Vol.67Issue(1) ：143-159.DOI:10.1007/s11433-023-2192-9

(DarkAI)Mapping the large-scale density field of dark matter using artificial intelligence

Zitong Wang Feng Shi Xiaohu Yang Qingyang Li Yanming Liu Xiaoping Li
中国科学:物理学 力学 天文学(英文版)2024，Vol.67Issue(1) ：143-159.DOI:10.1007/s11433-023-2192-9
✕
  • 维普
  • 万方数据

(DarkAI)Mapping the large-scale density field of dark matter using artificial intelligence

Zitong Wang 1Feng Shi 1Xiaohu Yang 2Qingyang Li 3Yanming Liu 1Xiaoping Li1
扫码查看

作者信息

  • 1. School of Aerospace Science and Technology,Xidian University,Xi'an 710126,China
  • 2. Department of Astronomy,School of Physics and Astronomy,and Shanghai Key Laboratory for Particle Physics and Cosmology,Shanghai Jiao Tong University,Shanghai 200240,China;Tsung-Dao Lee Institute and Key Laboratory for Particle Physics,Astrophysics and Cosmology,Ministry of Education,Shanghai Jiao Tong University,Shanghai 200240,China
  • 3. Department of Astronomy,School of Physics and Astronomy,and Shanghai Key Laboratory for Particle Physics and Cosmology,Shanghai Jiao Tong University,Shanghai 200240,China;Institute for Astronomy,University of Edinburgh,Royal Observatory,Edinburgh EH9 3HJ,United Kingdom
  • 折叠

Abstract

Herein,we present a deep-learning technique for reconstructing the dark-matter density field from the redshift-space distribution of dark-matter halos.We built a UNet-architecture neural network and trained it using the COmoving Lagrangian Acceleration fast simulation,which is an approximation of the N-body simulation with 5123 particles in a box size of 500h-1Mpc.Further,we tested the resulting UNet model not only with training-like test samples but also with standard N-body simulations,such as the Jiutian simulation with 61443 particles in a box size of 1000 h-1Mpc and the ELUCID simulation,which has a different cosmol-ogy.The real-space dark-matter density fields in the three simulations can be reconstructed reliably with only a small reduction of the cross-correlation power spectrum at 1％ and 10％ levels at k=0.1 and 0.3 h Mpc-1,respectively.The reconstruction clearly helps to correct for redshift-space distortions and is unaffected by the different cosmologies between the training(Planck2018)and test samples(WMAP5).Furthermore,we tested the application of the UNet-reconstructed density field to obtain the velocity & tidal field and found that this approach provides better results compared with the traditional approach based on the linear bias model,showing a 12.2％improvement in the correlation slope and a 21.1％reduction in the scatter between the predicted and true velocities.Thus,our method is highly efficient and has excellent extrapolation reliability beyond the training set.This provides an ideal solution for determining the three-dimensional underlying density field from the plentiful galaxy survey data.

Key words

dark matter/large-scale structure/cosmology

引用本文复制引用

基金项目

National SKA Program of China(2022SKA0110200)

National SKA Program of China(2022SKA0110202)

National Natural Science Foundation of China(12103037)

National Natural Science Foundation of China(11833005)

National Natural Science Foundation of China(11890692)

111 Project(B20019)

Shanghai Natural Science Foundation(19ZR1466800)

Science Research grants from the China Manned Space Project(CMS-CSST-2021-A02)

Fundamental Research Funds for the Central Universities(XJS221312)

High-Petformance Computing Platform of Xidian University()

出版年

2024
中国科学:物理学 力学 天文学(英文版)
中国科学院

中国科学:物理学 力学 天文学(英文版)

CSTPCD
影响因子：0.91
ISSN：1674-7348
参考文献量84
段落导航相关论文