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 2403.07498astro-ph.CO
收藏

FLAME:利用机器学习拟合Lyα吸收线

FLAME: Fitting Ly$α$ Absorption lines using Machine learning

Priyanka Jalan, Vikram Khaire, M. Vivek, Prakash Gaikwad

AI总结:

本研究提出机器学习算法FLAME,通过深度卷积神经网络拟合Lyα吸收线的Voigt轮廓,在模拟数据上精度优异,真实数据精度下降源于训练样本难以复刻真实数据的复杂性。

摘要:

我们推出了FLAME,这是一种利用深度卷积神经网络对HI莱曼阿尔法(Lyα)吸收线进行Voigt轮廓拟合的机器学习算法。FLAME整合了两种算法:第一种确定拟合Lyα吸收线所需的成分数量,第二种计算多普勒参数b、HI柱密度N_HI以及各成分的速度间隔。\n针对当前版本的FLAME,我们使用哈勃空间望远镜(HST)搭载的宇宙起源摄谱仪(COS)的远紫外光栅观测到的低红移Lyα森林对其进行训练。基于这些数据,我们用约10^6条模拟Voigt轮廓训练FLAME——这些轮廓通过正向建模模拟HST-COS观测到的Lyα吸收线,以此将谱线分类为单成分或双成分,进而确定Voigt轮廓拟合参数。\nFLAME在模拟数据上展现出出色的精度:识别出超过98%的单成分谱线和90%的双成分谱线。对于90%的单(双)成分谱线,其测得的b值误差在约±8(15)km s^-1以内,log N_HI/cm²值误差在约±0.3(0.8)以内。不过,应用于真实数据时,FLAME的成分分类精度下降了约10%。尽管如此,传统Voigt轮廓拟合方法得到的b和N_HI分布与FLAME的预测结果仍具有合理的一致性。\n我们设计了模拟真实数据参数的模拟HST-COS数据分析,结果表明FLAME能够达到与模拟数据表现相当的一致精度。这一发现说明,FLAME在真实数据上的精度下降,主要源于训练样本难以复刻真实数据的全部复杂性。

英文摘要:

We introduce FLAME, a machine-learning algorithm designed to fit Voigt profiles to HI Lyman-alpha (Ly$α$) absorption lines using deep convolutional neural networks. FLAME integrates two algorithms: the first determines the number of components required to fit Ly$α$ absorption lines, and the second calculates the Doppler parameter $b$, the HI column density N$_{\rm HI}$, and the velocity separation of individual components. For the current version of FLAME, we trained it on low-redshift Ly$α$ forests observed with the far-ultraviolet gratings of the Cosmic Origin Spectrograph (COS) on board the Hubble Space Telescope (HST). Using these data, we trained FLAME on $\sim$ $10^6$ simulated Voigt profiles which we forward-modeled to mimic Ly$α$ absorption lines observed with HST-COS in order to classify lines as either single or double components and then determine Voigt profile-fitting parameters. FLAME shows impressive accuracy on the simulated data, identifying more than 98\% (90\%) of single (double) component lines. It determines $b$ values within $\approx \pm{8}~(15)$ km s$^{-1}$ and log $N_{\rm HI}/ {\rm cm}^2$ values within $\approx \pm 0.3~(0.8)$ for 90\% of the single (double) component lines. However, when applied to real data, FLAME's component classification accuracy drops by $\sim$ 10\%. Nevertheless, there is reasonable agreement between the $b$ and N$_{\rm HI}$ distributions obtained from traditional Voigt profile-fitting methods and FLAME's predictions. Our mock HST-COS data analysis, designed to emulate real data parameters, demonstrates that FLAME is able to achieve consistent accuracy comparable to its performance with simulated data. This finding suggests that the drop in FLAME's accuracy when used on real data primarily arises from the difficulty in replicating the full complexity of real data in the training sample.

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。

↑