| Speaker: | Laurence Perreault Levasseur and Yashar Hezaveh (U Montreal) |
|---|---|
| Title: | Data-Driven High-Dimensional Inverse Problems: A Journey Through Strong Gravitational Lensing Data Analysis |
| Date (JST): | Thu, Jul 23, 2026, 13:30 - 15:00 |
| Place: | Seminar Room A |
| Abstract: |
Many inverse problems in astrophysics and cosmology do not admit analytic solutions or closed-form likelihoods. The talk concentrates on machine-learning solutions to this problem. In particular, learned priors over galaxy images in the context of strong gravitational lensing are discussed. Furthermore, simulation-based inference and new techniques for validation of neural posteriors are presented. |
