Anomaly Detection for Multivariate Time Series of Exotic Supernovae

TL;DR

This paper introduces an unsupervised, real-time anomaly detection pipeline for multivariate supernova time series using RNN-based autoencoders and isolation forests, capable of identifying unusual events in large astronomical datasets.

Contribution

It presents the first online anomaly detection method for supernovae, combining RNN autoencoders and isolation forests to identify unexpected phenomena in real-time data streams.

Findings

Successfully detected anomalous supernovae in simulated data

Identified objects with incorrect redshift measurements

Demonstrated effectiveness in real-time, online data streams

Abstract

Supernovae mark the explosive deaths of stars and enrich the cosmos with heavy elements. Future telescopes will discover thousands of new supernovae nightly, creating a need to flag astrophysically interesting events rapidly for followup study. Ideally, such an anomaly detection pipeline would be independent of our current knowledge and be sensitive to unexpected phenomena. Here we present an unsupervised method to search for anomalous time series in real time for transient, multivariate, and aperiodic signals. We use a RNN-based variational autoencoder to encode supernova time series and an isolation forest to search for anomalous events in the learned encoded space. We apply this method to a simulated dataset of 12,159 supernovae, successfully discovering anomalous supernovae and objects with catastrophically incorrect redshift measurements. This work is the first anomaly detection pipeline for supernovae which works with online datastreams.