Presentation #203.04 in the session “Machine Learning in Astronomy: Transient Discovery with Machine Learning (Meeting-in-a-Meeting)”.
The Vera C. Rubin Observatory Legacy Survey of Space and Time (LSST) will revolutionise transient astronomy, discovering millions of transients every night. Machine learning classification techniques will be critical to handle this data deluge. In this talk, I will touch on my work on classification but focus on the more challenging problem of finding rare and unexpected anomalies in this enormous dataset. I will introduce a new anomaly detection framework, Astronomaly, that can automatically sift through hundreds of thousands of light curves, picking out the most unusual objects. It also contains an active learning component to dynamically manage false positive artefacts and uninteresting anomalies.