Discovering Gravitationally Lensed Supernovae with Machine Learning in the LSST Era

The next generation of wide-field sky surveys will transform time-domain astrophysics. Among the most exciting prospects are gravitationally lensed supernovae—rare events that provide an independent probe of cosmology through time-delay and magnification measurements, while enabling unprecedented studies of high-redshift supernovae thanks to lensing amplification. However, identifying such events amid the enormous data stream expected from the Vera C. Rubin Observatory’s Legacy Survey of Space and Time (LSST), tens of millions of alerts every night, poses a major challenge. This project links the Oskar Klein Centre (OKC) in Stockholm with the Department of Physics and Astronomy at the University of the Western Cape (UWC), which is a member of the Inter-University Institute for Data-Intensive Astronomy (IDIA). The goal is to develop advanced machine-learning (ML) methods to detect lensed supernovae in real time. During a two-month stay at STIAS, I will collaborate closely with Dr. Michelle Lochner—whose group has pioneered active-learning and Bayesian ML approaches for transient discovery— to design and train ML architectures on simulated and archival survey data. The expected outcomes are a prototype ML classifier optimized for LSST alerts and a framework for sustained OKC–UWC/IDIA collaboration in data-intensive cosmology.