当前位置: X-MOL 学术Astrophys. J.  › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
Galaxy Merger Rates up to z ∼ 3 Using a Bayesian Deep Learning Model: A Major-merger Classifier Using IllustrisTNG Simulation Data
The Astrophysical Journal ( IF 4.9 ) Pub Date : 2020-06-03 , DOI: 10.3847/1538-4357/ab8f9b
Leonardo Ferreira 1 , Christopher J. Conselice 1 , Kenneth Duncan 2, 3 , Ting-Yun Cheng 1 , Alex Griffiths 1 , Amy Whitney 1
Affiliation  

Merging is potentially the dominate process in galaxy formation, yet there is still debate about its history over cosmic time. To address this we classify major mergers and measure galaxy merger rates up to z $\sim$ 3 in all five CANDELS fields (UDS, EGS, GOODS-S, GOODS-N, COSMOS) using deep learning convolutional neural networks (CNNs) trained with simulated galaxies from the IllustrisTNG cosmological simulation. The deep learning architecture used is objectively selected by a Bayesian Optmization process over the range of possible hyperparameters. We show that our model can achieve 90% accuracy when classifying mergers from the simulation, and has the additional feature of separating mergers before the infall of stellar masses from post mergers. We compare our machine learning classifications on CANDELS galaxies and compare with visual merger classifications from Kartaltepe et al. (2015), and show that they are broadly consistent. We finish by demonstrating that our model is capable of measuring galaxy merger rates, $\mathcal{R}$, that are consistent with results found for CANDELS galaxies using close pairs statistics, with $\mathcal{R}(z) = 0.02 \pm 0.004 \times (1 +z) ^ {2.76 \pm 0.21}$. This is the first general agreement between major mergers measured using pairs and structure at z < 3.

中文翻译:

银河合并率高达 z ∼ 3 使用贝叶斯深度学习模型:使用 IllustrisTNG 模拟数据的主要合并分类器

合并可能是星系形成的主要过程,但关于它在宇宙时间的历史仍然存在争议。为了解决这个问题,我们使用经过训练的深度学习卷积神经网络 (CNN) 在所有五个 CANDELS 场(UDS、EGS、GOODS-S、GOODS-N、COSMOS)中对主要合并进行分类并测量高达 z$\sim$3 的星系合并率使用来自 IllustrisTNG 宇宙学模拟的模拟星系。所使用的深度学习架构是通过贝叶斯优化过程在可能的超参数范围内客观选择的。我们表明,我们的模型在从模拟中对合并进行分类时可以达到 90% 的准确率,并且具有在恒星质量下降之前将合并与合并后分离的附加功能。我们比较了我们在 CANDELS 星系上的机器学习分类,并与 Kartaltepe 等人的视觉合并分类进行了比较。(2015),并表明它们大致上是一致的。最后,我们证明我们的模型能够测量星系合并率 $\mathcal{R}$,这与使用近对统计对 CANDELS 星系发现的结果一致,$\mathcal{R}(z) = 0.02 \ pm 0.004 \times (1 +z) ^ {2.76 \pm 0.21}$。这是使用 z < 3 处的对和结构测量的主要合并之间的第一个普遍协议。$\mathcal{R}(z) = 0.02 \pm 0.004 \times (1 +z) ^ {2.76 \pm 0.21}$。这是使用 z < 3 处的对和结构测量的主要合并之间的第一个普遍协议。$\mathcal{R}(z) = 0.02 \pm 0.004 \times (1 +z) ^ {2.76 \pm 0.21}$。这是使用 z < 3 处的对和结构测量的主要合并之间的第一个普遍协议。
更新日期:2020-06-03
down
wechat
bug