deepSIP：通过深度学习将Ia型超新星光谱与光度测量联系起来
deepSIP: Linking Type Ia Supernova Spectra to Photometric Quantities with Deep Learning
摘要：
我们提出了{ \tt deepSIP}（超新星Ia参数的深度学习），这是一个用于测量相位的软件包-首次使用深度学习-从光谱测量Ia型超新星（SN~Ia）的光变曲线形状。在其核心，{ \tt deepSIP}由三个卷积神经网络组成，这些神经网络在所有公开的低红移SN~Ia光谱的相当大一部分上进行训练，我们已经仔细地将光度学导出量耦合到这些光谱上。我们描述了我们的光谱和光度数据集的积累，为确保质量而采取的削减措施，以及我们用于拟合光变曲线的标准化技术。这些考虑产生了2754个光谱与光度特征的相位和光变曲线形状的汇编。虽然这样的样本在SN社区中很重要，但根据深度学习标准，它很小，因为网络通常有数百万甚至数十亿个自由参数。因此，我们引入了一种数据增强策略，该策略有意义地增加了我们分配用于训练的子集的大小，同时优先考虑模型鲁棒性和望远镜不可知性。我们通过将模型部署在训练和超参数选择过程中看不到的样本上来证明我们的模型的有效性，发现模型I识别具有$-10$和18\，d之间的相位和光曲线形状的光谱，由$Δm_{15}$参数化，在0.85和1.55\，mag之间，准确度为94.6\%。对于在相位- $Δm_{15}$空间中确实落在上述区域内的那些光谱，模型II预测相位的均方根误差（RMSE）为1.00 μ m，模型III预测$Δm_{15}$值的RMSE为0.068 μ m。
英文摘要：
We present {\tt deepSIP} (deep learning of Supernova Ia Parameters), a software package for measuring the phase and -- for the first time using deep learning -- the light-curve shape of a Type Ia supernova (SN~Ia) from an optical spectrum. At its core, {\tt deepSIP} consists of three convolutional neural networks trained on a substantial fraction of all publicly-available low-redshift SN~Ia optical spectra, onto which we have carefully coupled photometrically-derived quantities. We describe the accumulation of our spectroscopic and photometric datasets, the cuts taken to ensure quality, and our standardised technique for fitting light curves. These considerations yield a compilation of 2754 spectra with photometrically characterised phases and light-curve shapes. Though such a sample is significant in the SN community, it is small by deep-learning standards where networks routinely have millions or even billions of free parameters. We therefore introduce a data-augmentation strategy that meaningfully increases the size of the subset we allocate for training while prioritising model robustness and telescope agnosticism. We demonstrate the effectiveness of our models by deploying them on a sample unseen during training and hyperparameter selection, finding that Model~I identifies spectra that have a phase between $-10$ and 18\,d and light-curve shape, parameterised by $Δm_{15}$, between 0.85 and 1.55\,mag with an accuracy of 94.6\%. For those spectra that do fall within the aforementioned region in phase--$Δm_{15}$ space, Model~II predicts phases with a root-mean-square error (RMSE) of 1.00\,d and Model~III predicts $Δm_{15}$ values with an RMSE of 0.068\,mag.
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。
