arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 1908.10590astro-ph.COgr-qc
收藏

基于大尺度结构深度学习的宇宙学参数估计

Cosmological parameter estimation from large-scale structure deep learning

Shuyang Pan, Miaoxin Liu, Jaime Forero-Romero, Cristiano G. Sabiu, Zhigang Li, Haitao Miao, Xiao-Dong Li

摘要:

我们提出了一种轻量级的深度卷积神经网络(CNN),以高精度地从模拟的三维暗物质分布中估计宇宙学参数。训练集基于边长为$256\ h^{-1}\ \rm Mpc$的立方体框的465个实现,利用在$128^3$体素的立方体网格上内插的$128^3$粒子进行采样。这些体积具有在$0.16 \leq Ω_m \leq 0.46$和$2.0 \leq 10^9 A_s \leq 2.3$的平坦$Λ$ CDM参数空间内变化的宇宙学参数。神经网络以具有$32^3$体素的立方体作为输入,并且具有三个卷积层、三个密集层以及一些批归一化和池化层。在来自网络的最终预测中,我们发现原始振幅$σ_8$上的$2.5\%$偏差不能通过继续训练容易地解决。我们纠正这种偏差,以获得前所未有的准确性,在宇宙学参数估计与统计不确定性$δΩ_m$ =0.0015和$δσ_8$ =0.0029,这是几倍,比以前的CNN作品的结果。与使用0-130和10-130 $h^{-1}$ Mpc的聚类区域的2点分析方法相比,CNN约束分别精确数倍和一个数量级。最后,我们进行初步检查的神经网络的容错能力,并发现它具有鲁棒性对平滑,掩蔽,随机噪声,全局变化,旋转,反射和模拟分辨率。这些影响在典型的聚类分析中得到了很好的理解,但以前没有对CNN方法进行过测试。我们的工作表明,CNN可以比人们预期的更有希望从宇宙大尺度结构推导出严格的宇宙学约束。

英文摘要:

We propose a light-weight deep convolutional neural network (CNN) to estimate the cosmological parameters from simulated 3-dimensional dark matter distributions with high accuracy. The training set is based on 465 realizations of a cubic box with a side length of $256\ h^{-1}\ \rm Mpc$, sampled with $128^3$ particles interpolated over a cubic grid of $128^3$ voxels. These volumes have cosmological parameters varying within the flat $Λ$CDM parameter space of $0.16 \leq Ω_m \leq 0.46$ and $2.0 \leq 10^9 A_s \leq 2.3$. The neural network takes as an input cubes with $32^3$ voxels and has three convolution layers, three dense layers, together with some batch normalization and pooling layers. In the final predictions from the network we find a $2.5\%$ bias on the primordial amplitude $σ_8$ that can not easily be resolved by continued training. We correct this bias to obtain unprecedented accuracy in the cosmological parameter estimation with statistical uncertainties of $δΩ_m$=0.0015 and $δσ_8$=0.0029, which are several times better than the results of previous CNN works. Compared with a 2-point analysis method using clustering region of 0-130 and 10-130 $h^{-1}$ Mpc, the CNN constraints are several times and an order of magnitude more precise, respectively. Finally, we conduct preliminary checks of the error-tolerance abilities of the neural network, and find that it exhibits robustness against smoothing, masking, random noise, global variation, rotation, reflection, and simulation resolution. Those effects are well understood in typical clustering analysis, but had not been tested before for the CNN approach. Our work shows that CNN can be more promising than people expected in deriving tight cosmological constraints from the cosmic large scale structure.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑