In this paper, we address the problem of spectroscopic redshift estimation in Astronomy. Due to the expansion of the Universe, galaxies recede from each other on average. This movement causes the emitted electromagnetic waves to shift from the blue part of the spectrum to the red part, due to the Doppler effect. Redshift is one of the most important observables in Astronomy, allowing the measurement of galaxy distances. Several sources of noise render the estimation process far from trivial, especially in the low signal-to-noise regime of many astrophysical observations. In recent years, new approaches for a reliable and automated estimation methodology have been sought out, in order to minimize our reliance on currently popular techniques that heavily involve human intervention. The fulfilment of this task has evolved into a grave necessity, in conjunction with the insatiable generation of immense amounts of astronomical data. In our work, we introduce a novel approach based on Deep Convolutional Neural Networks. The proposed methodology is extensively evaluated on a spectroscopic dataset of full spectral energy galaxy distributions, modelled after the upcoming Euclid satellite galaxy survey. Experimental analysis on observations of idealistic and realistic conditions demonstrate the potent capabilities of the proposed scheme.
DeepDISC-Euclid: Source Classification and Photometric Redshifts in Euclid Deep Field North With a Pixel-Level Deep Learning Approach
University of Illinois at Urbana-Champaign(伊利诺伊大学厄巴纳-香槟分校);National Center for Supercomputing Applications, University of Illinois at Urbana-Champaign(美国国家超级计算应用中心,伊利诺伊大学厄巴纳-香槟分校);Center for Artificial Intelligence Innovation, University of Illinois at Urbana-Champaign(人工智能创新中心,伊利诺伊大学厄巴纳-香槟分校);Max-Planck-Institut für extraterrestrische Physik(马克斯·普朗克地外物理学研究所);Instituto de Astrofísica de Canarias (IAC)(加那利天体物理研究所;拉帕尔马大学天体物理学系);Departamento de Astrofísica, Universidad de La Laguna (ULL)(布里斯托大学物理学院);School of Physics, HH Wills Physics Laboratory, University of Bristol
Predicting the Spectroscopic Features of Galaxies by Applying Manifold Learning on Their Broad-Band Colors: Proof of Concept and Potential Applications for Euclid, Roman, and Rubin LSST
The Spectroscopic Binaries from LAMOST Medium-Resolution Survey (MRS). I. Searching for Double-lined Spectroscopic Binaries (SB2s) with Convolutional Neural Network
European Southern Observatory(欧洲南方天文台);Institute of Physics, École Polytechnique Fédérale de Lausanne (EPFL)(洛桑联邦理工学院物理研究所);IRFU, CEA, Université Paris-Saclay(法国原子能和替代能源署法国研究和技术核物理研究所巴黎萨克雷大学);Department of Astronomy, Faculty of Mathematics, University of Belgrade(贝尔格莱德大学数学学院天文学系);Hamburger Sternwarte, Universität Hamburg(汉堡大学天文台);INAF OACN(意大利国家天体物理研究所卡塔尼亚天文台);Astronomical Institute of Charles University(查理大学天文研究所);Instituto de Astrofísica de Canarias(加那利群岛天体物理研究所)
cloelike: A Python Library for Cosmological Likelihood Inference in the Euclid Era
Waterloo Centre for Astrophysics, University of Waterloo(滑铁卢天体物理中心,滑铁卢大学);Leiden Observatory, Leiden University(莱顿天文台,莱顿大学);Centre for Astrophysics Research, University of Hertfordshire(天体物理学研究中心,赫特福德大学);Scientific Information, German Aerospace Center (DLR)(科学信息部,德国航空航天中心);INAF - Osservatorio Astronomico di Trieste(意大利国家天体物理研究所 - 的里雅斯特天文台);INAF - Osservatorio Astronomico di Brera(意大利国家天体物理研究所 - 布雷拉天文台);Institute of Space Sciences (ICE, CSIC)(空间科学研究所);INAF - IASF Milano(意大利国家天体物理研究所 - 米兰天体物理研究所);Astronomical Institute (AIRUB), Ruhr University Bochum(天文研究所,鲁尔大学波鸿分校);Department of Physics, University of Oxford(物理系,牛津大学);Dipartimento di Fisica "Aldo Pontremoli", Università degli Studi di Milano(物理系,米兰大学);IRAP, Université de Toulouse(天体物理研究所,图卢兹大学);Osservatorio Astronomico di Roma(罗马天文台);Mullard Space Science Laboratory, University College London(穆拉德太空科学实验室,伦敦大学学院);Software Institut d’Estudis Espacials de Catalunya (IEEC)(加泰罗尼亚空间研究研究所)
$\texttt{BTSbot}$: A Multi-input Convolutional Neural Network to Automate and Expedite Bright Transient Identification for the Zwicky Transient Facility
Enhancing Lyα Emitter Identification in HETDEX with a Convolutional Neural Network
The University of Texas at Austin(德克萨斯大学奥斯汀分校);MIRAI Technology Institute, Shiseido Co., Ltd.(资生堂株式会社 MIRAI 技术研究所);McDonald Observatory, The University of Texas at Austin(德克萨斯大学奥斯汀分校麦克唐纳天文台);The University of Chicago(芝加哥大学);The SkAI Institute(SkAI 研究所)
The RAdial Velocity Experiment (RAVE): Parameterisation of RAVE spectra based on convolutional neural networks
在对恒星进行大型光谱调查的背景下,数据驱动的方法是在短时间内推导数百万光谱的物理参数的关键。卷积神经网络(CNN)使我们能够将观测值(例如光谱,恒星星等)与物理属性(大气参数,化学丰度或一般标签)联系起来。我们训练了一个CNN,采用来自APOGEE DR 16(分辨率R=22500)数据的恒星大气参数和化学丰度作为训练集标签。作为输入,我们使用了与APOGEE DR 16数据重叠的部分中等分辨率RAVE DR 6光谱(R~7500)以及宽带ALL_WISE和2 MASS测光,以及Gaia DR 2测光和反射。我们推导出精确的大气参数Teff、log(g)和[M/H]以及420165 RAVE光谱的[Fe/H]、[alpha/M]、[Mg/Fe]、[Si/Fe]、[Al/Fe]和[Ni/Fe]的化学丰度。精度通常为60 K(Teff),0.06 log(g)和0.02-0.04 dex(单个化学丰度)。将测光和天体测量作为额外的约束条件,大大提高了结果的准确性和精度的衍生标签。我们提供了一个目录的CNN训练的大气参数和丰度以及它们的不确定性为420165星在RAVE调查。基于CNN的方法提供了一种强大的方法来结合光谱,光度和天体测量数据,而不需要以恒星演化模型的形式应用任何先验。开发的程序可以将RAVE光谱的科学输出扩展到DR 6之外,以进行和计划进行的调查,如Gaia RVS,4 …
Fanaroff-Riley classification of radio galaxies using group-equivariant convolutional neural networks
卷积神经网络(CNN)中的权重共享确保了它们的特征映射将是对称等变的。然而,尽管常规卷积对于平移是等变的,但是它们对于输入图像数据的其他等距(诸如旋转和反射)不是等变的。对于天文对象的分类,如射电星系,这是统计学上预计是全球方向不变的,这种缺乏二面角等方差意味着传统的CNN必须明确学习分类一个特定类型的对象的所有旋转版本单独。在这项工作中,我们首次将群等变卷积神经网络应用于射电星系分类,并通过保留包含平移、旋转和反射的欧几里得群E(2)的等方差来探索它们降低类内变异性的潜力。对于这里考虑的射电星系分类问题,我们发现,分类性能适度提高使用循环和二面角模型,而无需额外的超参数调整,和D16等变模型提供了最好的测试性能。我们使用Monte Carlo Dropout方法作为贝叶斯近似来恢复作为图像方向的函数的认知不确定性,并表明E(2)-等变模型能够减少作为旋转函数的模型置信度的变化。