Abstract:New time-domain surveys, such as the Vera C. Rubin Observatory Legacy Survey of Space and Time (LSST), will observe millions of transient alerts each night, making standard approaches of visually identifying new and interesting transients infeasible. We present two novel methods of automatically detecting anomalous transient light curves in real-time. Both methods are based on the simple idea that if the light curves from a known population of transients can be accurately modelled, any deviations from model predictions are likely anomalies. The first modelling approach is a probabilistic neural network built using Temporal Convolutional Networks (TCNs) and the second is an interpretable Bayesian parametric model of a transient. We demonstrate our methods' ability to provide anomaly scores as a function of time on light curves from the Zwicky Transient Facility. We show that the flexibility of neural networks, the attribute that makes them such a powerful tool for many regression tasks, is what makes them less suitable for anomaly detection when compared with our parametric model. The parametric model is able to identify anomalies with respect to common supernova classes with high precision and recall scores, achieving area under the precision-recall curves (AUCPR) above 0.79 for most rare classes such as kilonovae, tidal disruption events, intermediate luminosity transients, and pair-instability supernovae. Our ability to identify anomalies improves over the lifetime of the light curves. Our framework, used in conjunction with transient classifiers, will enable fast and prioritised followup of unusual transients from new large-scale surveys.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is to detect anomalies in real - time during large - scale transient sky surveys. With the launch of new time - domain survey projects such as the Vera C. Rubin Observatory's Legacy Survey of Space and Time (LSST), millions of transient alerts may be observed every night, which makes it impractical to discover new and interesting transient phenomena through traditional methods (such as manual visual identification). Therefore, this paper proposes two novel methods for automatically detecting anomalies in transient light curves. These methods can be immediately processed and analyzed when the data stream arrives in real - time, thereby quickly identifying rare or unknown types of transient events. Specifically, the research aims to: 1. **Develop real - time anomaly detection methods**: Develop new methods capable of detecting abnormal transient light curves in real - time for the large amount of data generated in large - scale transient sky surveys. 2. **Improve the accuracy of anomaly detection**: Improve the precision and recall rate of anomaly detection by using two different modeling methods - a probabilistic neural network based on Temporal Convolutional Networks (TCNs) and an interpretable Bayesian parameter model. 3. **Adapt to different types of transient events**: Ensure that the proposed methods can effectively identify various known transient types (such as supernovae, tidal disruption events, etc.) as well as rare or unknown types of transient events. 4. **Optimize the early identification ability**: Identify anomalies in the early stages of transient events so that subsequent detailed observations can be carried out in a timely manner, which is crucial for understanding the physical mechanisms of transient events. Through these efforts, this research aims to provide astronomers with a powerful tool to cope with the upcoming big - data era, especially when processing large - scale transient sky survey data, enabling rapid screening of anomalies with scientific value, thereby promoting the development of astronomy.

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

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

A Classifier-Based Approach to Multi-Class Anomaly Detection for Astronomical Transients

Predicting the Age of Astronomical Transients from Real-Time Multivariate Time Series

Anomaly Detection and Approximate Similarity Searches of Transients in Real-time Data Streams

NEural Engine for Discovering Luminous Events ( NEEDLE ): identifying rare transient candidates in real time from host galaxy images

Detecting optical transients using artificial neural networks and reference images from different surveys

NEural Engine for Discovering Luminous Events (NEEDLE): identifying rare transient candidates in real time from host galaxy images

A Classifier-Based Approach to Multi-Class Anomaly Detection Applied to Astronomical Time-Series

RAPID: Early Classification of Explosive Transients using Deep Learning

Transient-optimised real-bogus classification with Bayesian Convolutional Neural Networks -- sifting the GOTO candidate stream

Finding radio transients with anomaly detection and active learning based on volunteer classifications

Understanding of the properties of neural network approaches for transient light curve approximations

Nonparametric Transient Classification using Adaptive Wavelets

A transient search using combined human and machine classifications

Real-Time Detection of Unmodelled Gravitational-Wave Transients Using Convolutional Neural Networks

Transient Classification in low SNR Gravitational Wave data using Deep Learning

Real-time, Value-driven Data Augmentation in the Era of LSST

TransientViT: A novel CNN - Vision Transformer hybrid real/bogus transient classifier for the Kilodegree Automatic Transient Survey

Real-time abnormal detection of GWAC light curve based on wavelettransform combined with GRU-Attention