Galaxies SIG Seminar, 8 Sept 2026
Learning to See Sharper: Deep Learning for Astronomical Data Enhancement. Speaker: Shoubaneh Hemmati The post Galaxies SIG Seminar, 8 Sept 2026 appeared first on NASA Science .
The Cosmic Origins Community held its Galactic Science Interest Group Seminar on September 8, 2026. The event was conducted virtually at 1:00 PM ET. Dr. Shoubaneh Hemmati from Caltech/IPAC presented her research titled "Learning to See Sharper: Deep Learning for Astronomical Data Enhancement."
Dr. Hemmati explained that astronomical surveys often face trade-offs between coverage, depth, and resolution, creating an ill-posed inverse problem. Classical deconvolution methods often struggle to overcome these challenges. Her work aimed to address this issue using deep learning techniques.
She discussed specialized neural architectures designed to tackle various imaging and spectroscopic challenges. These included generative models for deblending crowded fields and super-resolving infrared observations, shape-preserving residual networks for weak lensing, diffusion models for realistic survey simulations, and physics-informed models that resolved blended spectral emission lines.
Dr. Hemmati emphasized that selecting the appropriate architecture for specific scientific measurements could significantly enhance data enhancement across next-generation surveys. The seminar provided insights into how deep learning could be leveraged to maximize the potential of astronomical surveys.
Written by urgent.news from NASA Science's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
