A Deep Learning Approach to Galaxy Cluster X-Ray Masses

被引：68

作者：

Ntampaka, M. ^{[1
,2
]}

ZuHone, J. ^{[3
]}

Eisenstein, D. ^{[1
]}

Nagai, D. ^{[4
]}

Vikhlinin, A. ^{[1
,5
]}

Hernquist, L. ^{[1
]}

Marinacci, F. ^{[1
]}

Nelson, D. ^{[6
]}

Pakmor, R. ^{[6
]}

Pillepich, A. ^{[7
]}

Torrey, P. ^{[8
]}

Vogelsberger, M. ^{[9
]}

机构：

[1] Harvard Smithsonian Ctr Astrophys, 60 Garden St, Cambridge, MA 02138 USA

[2] Harvard Univ, Harvard Data Sci Initiat, Cambridge, MA 02138 USA

[3] Smithsonian Astrophys Observ, Cambridge, MA 02138 USA

[4] Yale Univ, Dept Phys, New Haven, CT 06520 USA

[5] Space Res Inst IKI, Profsoyuznaya 84-32, Moscow, Russia

[6] Max Planck Inst Astrophys, Karl Schwarzschild Str 1, D-85741 Garching, Germany

[7] Max Planck Inst Astron, Konigstuhl 17, D-69117 Heidelberg, Germany

[8] Univ Florida, Dept Astron, 211 Bryant Space Sci Ctr, Gainesville, FL 32611 USA

[9] MIT, Kavli Inst Astrophys & Space Res, 77 Massachusetts Ave, Cambridge, MA 02139 USA

来源：

ASTROPHYSICAL JOURNAL | 2019年 / 876卷 / 01期

关键词：

galaxies: clusters: general; methods: statistical; X-rays: galaxies: clusters; LARGE-SCALE STRUCTURE; ILLUSTRISTNG SIMULATIONS; NEURAL-NETWORKS; COSMOLOGICAL SIMULATIONS; OBJECT RECOGNITION; EVOLUTION; SCATTER; TEMPERATURE; LUMINOSITY; PROJECT;

D O I：

10.3847/1538-4357/ab14eb

中图分类号：

P1 [天文学];

学科分类号：

0704 ;

摘要：

We present a machine-learning (ML) approach for estimating galaxy cluster masses from Chandra mock images. We utilize a Convolutional Neural Network (CNN), a deep ML tool commonly used in image recognition tasks. The CNN is trained and tested on our sample of 7896 Chandra X-ray mock observations, which are based on 329 massive clusters from the IllustrisTNG simulation. Our CNN learns from a low resolution spatial distribution of photon counts and does not use spectral information. Despite our simplifying assumption to neglect spectral information, the resulting mass values estimated by the CNN exhibit small bias in comparison to the true masses of the simulated clusters (-0.02 dex) and reproduce the cluster masses with low intrinsic scatter, 8% in our best fold and 12% averaging over all. In contrast, a more standard core-excised luminosity method achieves 15%-18% scatter. We interpret the results with an approach inspired by Google DeepDream and find that the CNN ignores the central regions of clusters, which are known to have high scatter with mass.

引用

页数：7

共 76 条

[21] Identifying reionization sources from 21 cm maps using Convolutional Neural Networks [J].