Three CAMELS papers posted to arXiv this week!  We perform multifield cosmology with artificial intelligence for the first time, show that neural networks can yield robust marginalization of baryonic effects for cosmological inference at the field level, and make publicly available the largest dataset of 2D maps and 3D grids with properties of cosmic gas, dark matter, and stars to train machine learning algorithms in the CAMELS Multifield Dataset.

Listen to Daniel Anglés-Alcázar’s interview with CBC radio host Bob McDonald for the award-winning radio science program Quirks & Quarks of the Canadian Broadcasting Corporation:  How does a black hole eat?