We will have a special AstroParticle Lunch seminar this Friday. We are excited to host Prof. Hy Trac from Carnegie Mellon University. He will be giving a 45-minute talk, followed by time for questions and discussions.
Location: Smith Seminar Room (1080 PRB) (note the different location than usual)
Time: 11:30 AM (ET), Friday, March 7
Title: Learning to See the Dark Matter in Galaxy Clusters
Abstract: Galaxy clusters contain vast amounts of cold dark matter, hot ionized gas, and tens to hundreds of visible galaxies. In 1933, Fritz Zwicky first proposed the existence of dark matter after observing the motions of galaxies in the Coma Cluster. Today, we estimate that dark matter makes up about 85% of the total matter in the Universe. However, the spatial distributions of dark matter and baryons within galaxy clusters remain uncertain. To address this, we first present an updated mass estimate of the Coma cluster using modern AI/ML techniques. We then show how generative diffusion models, trained on multi-wavelength images (e.g. SZ effect, X-ray emission, gravitational lensing), can predict the gas, dark matter, and total matter projected density fields. When applied to synthetic images of simulated clusters, these models produce accurate and unbiased mass maps. AI/ML offers a promising approach for mapping the unseen in galaxy clusters. Knowing where is the dark matter will help us to understand its nature and that of the Universe.