A Classification Algorithm for Time-domain Novelties in Preparation for LSST Alerts. Application to Variable Stars and Transients Detected with DECam in the Galactic Bulge

Abstract

With the advent of the Large Synoptic Survey Telescope (LSST), time-domain astronomy will be faced with an unprecedented volume and rate of data. Real-time processing of variables and transients detected by such large-scale surveys is critical to identifying the more unusual events and allocating scarce follow-up resources efficiently. We develop an algorithm to identify these novel events within a given population of variable sources. We determine the distributions of magnitude changes (dm) over time intervals (dt) for a given passband f , p f (dm|dt), and use these distributions to compute the likelihood of a test source being consistent with the population, or an outlier. We demonstrate our algorithm by applying it to the DECam multi-band time-series data of more than 2000 variable stars identified by Saha et al. (2019) in the Galactic Bulge that are largely dominated by long-period variables and pulsating stars. Our algorithm discovers 18 outlier sources in the sample, including a microlensing event, a dwarf nova, and two chromospherically active RS CVn stars, as well as sources in the Blue Horizontal Branch region of the color-magnitude diagram without any known counterparts. We compare the performance of our algorithm for novelty detection with multivariate KDE and Isolation Forest on the simulated PLAsTiCC dataset. We find that our algorithm yields comparable results despite its simplicity. Our method provides an efficient way for flagging the most unusual events in a real-time alert-broker system.

Real-time Detection of Anomalies in Multivariate Time Series of Astronomical Data

“…This work and other recent approaches to transient anomaly detection [e.g. 16,18,[20][21][22][23] are going to be critical for discovery in the new era of large-scale astronomical surveys.…”

Preparing to discover the unknown with Rubin LSST -- I: Time domain

“…Notable advances in the field have been achieved in recent years across disciplines: from threat detection in defense and security (e.g. Sultani et al 2018), to astrophysics (Soraisam et al 2020;Pruzhinskaya et al 2019;Ishida et al 2019;Aleo et al 2020;Vafaei Sadr et al 2019;Martínez-Galarza et al 2020;Lochner & Bassett 2020;Doorenbos et al 2020) with the discovery of rare and possible unique astrophysical phenomena (Lintott et al 2009;Micheli et al 2018;Boyajian et al 2018, although we note that two of these "true novelties" were detected through crowd-sourced data analysis). Anomaly detection is generally approached either through unsupervised or supervised learning learning techniques (e.g.…”

“…Yet one of the most exciting promises of LSST is its potential to discover completely novel phenomena, never before observed or predicted from theory. We created a five-fold F oM that relies on a set of MAFs that assesses the ability of Rubin Observatory LSST to discover novel astrophysical objects, but instead of selecting known anomalies (e.g., Boyajian et al 2018) or theoretically predicted unusual phenomena to benchmark our results, as more commonly done in the field (Soraisam et al 2020;Pruzhinskaya et 12. While the footprint family has the best performing OpSim overall for the Galactic Plane (footprint gp smooth) it also has the least performing OpSims, showing the largest dynamical range due to the combined effects of both the footprint and time-gap metrics.…”

Real-Time Detection of Anomalies in Large-Scale Transient Surveys

“…Recently, there have been a few anomaly detection algorithms applied to astronomical light curves (e.g. Rebbapragada et al 2009;Nun et al 2014;Solarz et al 2017;Giles & Walkowicz 2019;Sadeh 2019;Pruzhinskaya et al 2019;Ishida et al 2019;Soraisam et al 2020;Webb et al 2020;Villar et al 2020Villar et al , 2021Malanchev et al 2020;Martínez-Galarza et al 2020;Lochner & Bassett 2020). These approaches predominantly use unsupervised clustering algorithms such as Density-Based Spatial Clustering of Applications with Noise (DBSCAN) (e.g.…”

“…These approaches are effective at identifying anomalies once the full light curve has been observed, but many of them prove problematic for real-time detection in large-scale transient surveys. However, Soraisam et al (2020) and Villar et al (2021) have recently developed some of the first methods that perform real-time anomaly detection. Villar et al (2021) uses a variational recurrent autoencoder to learn an encoded form of each light curve before obtaining anomaly scores by passing the encoded form into an isolation forest.…”