Over the past decade wide-field optical time-domain surveys have increased the discovery rate of transients to the point that $\lesssim 10\%$ are being spectroscopically classified. Despite this, these surveys have enabled the discovery of new and rare types of transients, most notably the class of hydrogen-poor superluminous supernovae (SLSN-I), with about 150 events confirmed to date. Here we present a machine-learning classification algorithm targeted at rapid identification of a pure sample of SLSN-I to enable spectroscopic and multi-wavelength follow-up. This algorithm is part of the FLEET (Finding Luminous and Exotic Extragalactic Transients) observational strategy. It utilizes both light curve and contextual information, but without the need for a redshift, to assign each newly-discovered transient a probability of being a SLSN-I. This classifier can achieve a maximum purity of about 85\% (with 20\% completeness) when observing a selection of SLSN-I candidates. Additionally, we present two alternative classifiers that use either redshifts or complete light curves and can achieve an even higher purity and completeness. At the current discovery rate, the FLEET algorithm can provide about $20$ SLSN-I candidates per year for spectroscopic follow-up with 85\% purity; with the Legacy Survey of Space and Time we anticipate this will rise to more than $\sim 10^3$ events per year.

中文翻译：

---

FLEET：一种与 Redshift 无关的机器学习管道，可快速识别贫氢超亮超新星

在过去十年中，宽视场光时域调查已将瞬变的发现率提高到 $\lesssim 10\%$ 正在被光谱分类的程度。尽管如此，这些调查还是发现了新的和罕见的瞬变类型，最显着的是贫氢超光速超新星 (SLSN-I) 类，迄今为止已确认了大约 150 个事件。在这里，我们提出了一种机器学习分类算法，旨在快速识别 SLSN-I 的纯样本，以实现光谱和多波长跟踪。该算法是 FLEET（寻找发光和奇异的河外瞬变）观测策略的一部分。它利用光曲线和上下文信息，但不需要红移，为每个新发现的瞬态分配一个成为 SLSN-I 的概率。当观察选择的 SLSN-I 候选时，该分类器可以达到大约 85% 的最大纯度（具有 20% 的完整性）。此外，我们提出了两种替代分类器，它们使用红移或完整的光变曲线，可以实现更高的纯度和完整性。以目前的发现率，FLEET 算法每年可以提供约 20 美元的 SLSN-I 候选物用于光谱随访，纯度为 85%；通过《空间和时间的遗产调查》，我们预计这将增加到每年超过 $\sim 10^3$ 的事件。以目前的发现率，FLEET 算法每年可以提供约 20 美元的 SLSN-I 候选物用于光谱随访，纯度为 85%；通过《空间和时间的遗产调查》，我们预计这将增加到每年超过 $\sim 10^3$ 的事件。以目前的发现率，FLEET 算法每年可以提供约 20 美元的 SLSN-I 候选物用于光谱随访，纯度为 85%；通过《空间和时间的遗产调查》，我们预计这将增加到每年超过 $\sim 10^3$ 的事件。