HOLISMOKES——XII. 使用长短期记忆网络测量强引力透镜Ia型超新星的时间延迟
HOLISMOKES -- XII. Time-delay Measurements of Strongly Lensed Type Ia Supernovae using a Long Short-Term Memory Network
AI总结：
本文提出结合LSTM与全连接网络的LSTM-FCNN，用于跨观测模式测量强透镜Ia型超新星时间延迟，精度约0.7天且性能约为随机森林三倍。
摘要：
强引力透镜Ia型超新星（LSNe Ia）是直接测量哈勃常数（$H_0$）的一种有前景的探针。要将LSNe Ia用于宇宙学，需要测量多个图像之间的时间延迟、建立透镜质量模型，并沿视线方向进行质量重建。在本工作中，我们提出了机器学习网络LSTM-FCNN，它结合了长短期记忆网络（LSTM）和全连接神经网络（FCNN）。LSTM-FCNN旨在对涵盖广泛属性范围的LSNe Ia样本测量时间延迟；我们预计将在即将到来的鲁宾天文台空间和时间遗产调查（LSST）中发现这些超新星，并已为其计划后续观测。利用$i$波段的后续观测（观测节奏为1至3天，单历元5σ深度为24.5星等），我们在大量LSNe Ia样本上实现了无偏差的延迟测量，精度约为0.7天。LSTM-FCNN比随机森林（RF）等以往的机器学习方法通用得多：RF必须针对每种观测模式分别训练，而LSTM-FCNN的性能仍比RF高出约三倍。因此，LSTM-FCNN是获得LSNe Ia稳健时间延迟的一种非常有前景的方法，这对于精确、准确地约束$H_0$具有重要意义。
英文摘要：
Strongly lensed Type Ia supernovae (LSNe Ia) are a promising probe to measure the Hubble constant ($H_0$) directly. To use LSNe Ia for cosmography, a time-delay measurement between the multiple images, a lens-mass model, and a mass reconstruction along the line of sight are required. In this work, we present the machine learning network LSTM-FCNN which is a combination of a Long Short-Term Memory Network (LSTM) and a fully-connected neural network (FCNN). The LSTM-FCNN is designed to measure time delays on a sample of LSNe Ia spanning a broad range of properties, which we expect to find with the upcoming Rubin Observatory Legacy Survey of Space and Time (LSST) and for which follow-up observations are planned. With follow-up observations in $i$ band (cadence of one to three days with a single-epoch $5σ$ depth of 24.5 mag), we reach a bias-free delay measurement with a precision around 0.7 days over a large sample of LSNe Ia. The LSTM-FCNN is far more general than previous machine learning approaches such as the Random Forest (RF), where a RF has to be trained for each observational pattern separately, and yet the LSTM-FCNN outperforms the RF by a factor of roughly three. Therefore, the LSTM-FCNN is a very promising approach to achieve robust time delays in LSNe Ia, which is important for a precise and accurate constraint on $H_0$
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。
