Introduction
This bi-monthly seminar series explores real-world applications of physics-informed machine learning (Φ-ML) methods to the engineering practice. They cover a wide range of topics, offering a cross-sectional view of the state of the art on Φ-ML research, worldwide.
Participants have the opportunity to hear from leading researchers and learn about the latest developments in this emerging field. These seminars also offer the chance to identify and spark collaboration opportunities.
About the event
Strong gravitational lensing offers a powerful tool to probe the universe, from mapping the distribution of dark matter to studying the formation of distant galaxies. However, due to the nonlinear, non-convex, and very high-dimensional nature of the inverse problem, traditional lensing analysis methods have struggled to convincingly extract all available information from existing observations. This talk explores recent advancements in Bayesian image reconstruction and high-dimensional inference, with a focus on applications to strong gravitational lensing. Using diffusion models and score-based generative approaches, I will present methods for reconstructing high-fidelity astronomical images while addressing key challenges such as out-of-distribution data robustness, uncertainty quantification, and joint inference of hierarchical properties. I will also present statistical tools to assess the accuracy of posterior samples obtained with machine learning. These advances in deep learning-based inference open new possibilities for studying complex lensing systems at scale, particularly in the era of large surveys like LSST, Euclid, and the Roman Space Telescope.
