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 2604.12414astro-ph.GAastro-ph.IM
收藏

利用卷积神经网络增强HETDEX中的Lyα发射体识别

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)
Shiro Mukae, Erin Mentuch Cooper, Karl Gebhardt, Dustin Davis, Lindsay R. House, Mahdi Qezlou, Julian B. Muñoz, Shun Saito, Daniel J. Farrow, Caryl Gronwall, Do… 展开作者

Shiro Mukae, Erin Mentuch Cooper, Karl Gebhardt, Dustin Davis, Lindsay R. House, Mahdi Qezlou, Julian B. Muñoz, Shun Saito, Daniel J. Farrow, Caryl Gronwall, Donald P. Schneider, Eric Gawiser

AI总结:

本文提出一种深度学习框架,通过卷积神经网络提升HETDEX中低信噪比Lyα发射体的识别精度,实验结果显示在不同信噪比下模型均表现出色,有效减少误判并提高识别效率。

摘要:

我们提出一种深度学习框架,用于增强Hobby-Eberly Telescope Dark Energy Experiment (HETDEX)中Lyα发射体(LAEs)的识别,HETDEX是一项针对z在1.9到3.5之间的LAEs的无成像预选光谱学调查。我们主要解决低信噪比(S/N)区域(4.8≤S/N≤5.5),其中LAE候选者受到显著的噪声污染。为了区分LAE候选者与伪影和天空残差,我们采用了一个基于二维光谱图像的单发射线训练的卷积神经网络(CNN)。训练样本来自HETDEX COSMOS目录,外部验证来自辅助观测和我们的参与科学项目Dark Energy Explorers。对于小格式、低分辨率光谱数据,模型在高S/N区域(S/N>5.5)实现了平衡的准确率、精确率和召回率分别为94.1%、97.5%和97.5%,在低S/N区域分别为85.1%、78.2%和84.4%。使用HETDEX中独立由DESI光谱学识别的LAEs,模型恢复了99%和93%的高和低S/N LAEs。视觉归因表明,CNN在真阳性中关注平滑、空间扩展的中心发射,在真阴性中关注不规则或噪声特征。应用于完整的HETDEX目录,CNN将信噪比阈值降低到4.8,通过抑制z≈1.9-2.5红移分布中的虚假尖峰。我们的方法通过减少星系聚类中的假阳性,促进了HETDEX宇宙学分析,并突显了领域特定深度学习在无目标调查中精炼低信噪比光谱识别中的价值。

英文摘要:

We present a deep learning framework to enhance the identification of Ly$α$ emitters (LAEs) in the Hobby-Eberly Telescope Dark Energy Experiment (HETDEX), an untargeted spectroscopic survey of LAEs at $1.9 < z < 3.5$ without imaging pre-selection. We primarily address the low signal-to-noise ratio (S/N) regime ($4.8 \leq \mathrm{S/N} \leq 5.5$), where LAE candidates suffer from substantial noise contamination. To distinguish LAE candidates from artifacts and sky residuals, we employ a convolutional neural network (CNN) trained on two-dimensional spectral images of single emission lines. The training sample is constructed from the HETDEX COSMOS catalog, with external validation from ancillary observations and our participatory science project, \textit{Dark Energy Explorers}. For small-format, low-resolution spectroscopic data, the model achieves a balanced accuracy, precision, and recall of $94.1\%$, $97.5\%$, and $97.5\%$, respectively, in the high-S/N regime ($\mathrm{S/N}>5.5$), and $85.1\%$, $78.2\%$, and $84.4\%$ in the low-S/N regime. Using HETDEX LAEs independently identified by DESI spectroscopy, the model recovers $99\%$ and $93\%$ of the high- and low-S/N LAEs, respectively. Visual attribution indicates that the CNN attends to smooth, spatially extended central emission in true positives and to irregular or noisy features in true negatives. Applied to the full HETDEX catalog, the CNN enables an S/N threshold down to 4.8 by suppressing spurious spikes across $z\sim 1.9$--$2.5$ in the redshift distribution. Our approach facilitates HETDEX cosmological analyses by mitigating false positives in galaxy clustering and highlights the value of domain-specific deep learning for refining low-S/N spectroscopic identification in untargeted surveys.

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。

↑