CDS&E: Application of Machine Learning to Supermassive Black Hole Discovery with LSST

Next-generation astronomical surveys such as the NSF/DOE Vera C. Rubin Observatory's Legacy Survey of Space and Time (LSST) will observe billions of celestial objects, creating unprecedented opportunities to study the universe while presenting major challenges due to irregular observations, diverse data types, and limited training data. This project combines astronomy and data science to develop advanced AI and machine-learning methods that identify active galactic nuclei (AGN), or actively growing supermassive black holes, by integrating multiple types of observational data into a single probabilistic model. It also uses physically motivated models of AGN brightness variations to better understand black hole properties, including mass, accretion rate, and evolutionary stage. Together, these approaches will provide an accurate, interpretable, and scalable framework for AGN discovery in the LSST era, while producing publicly available catalogs, open-source software, and reproducible tools that will support future astronomical research and improve our understanding of black hole evolution. All code, catalogs, and workflows will be released publicly (via GitHub, Zenodo, and Docker) to enable community use and reproducibility. This project will unify heterogeneous diagnostics into a single probabilistic latent-space model, advancing an unbiased census of accreting black holes in the LSST era. Rather than optimizing separate diagnostics, the latent-space approach used in this project integrates all available information into a common probabilistic framework, transforming a traditionally fragmented problem into a coherent inference model. This research demonstrates how AI and machine-learning representations can be physically grounded while providing robust uncertainty quantification at the petabyte scale. Through this project, graduate and undergraduate students from both physics and information science will gain cross-disciplinary experience in astronomy and informatics, with training in machine learning, data management, and scientific communication. Outreach activities will be conducted through Drexel's Lynch Observatory and the Dornsife Center for Neighborhood Partnerships, creating learning opportunities open to all Philadelphia residents. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.