Project 7: AI Techniques for Inverse Problems in Cosmology

Lead PIs: Chihway Chang (UC), Rebecca Willett (UC)
Co-PIs: Joshua Frieman (UC), Moritz Münchmeyer (UW–Madison), Daniel Sanz-Alonso (UC)
Collaborators: Shrihan Agarwal (UC), Dhayaa Anbajagane (UC), Scott Dodelson (UC), Chun-Hao To (UC), Yuuki Omori (UC), Benjamin Remy (Princeton University), Georgios Valogiannis (UC), Shi-Hui Zang (UW–Madison), Roy Zhao (UC), Alan Zhou (UC) 
Project Summary: Inverse problems are concerned with estimating model parameters from indirect, noisy measurements. This project will develop novel artificial intelligence (AI) and data science techniques to solve a key inverse problem in cosmology: reconstructing the initial conditions of the universe from present-day measurements. Our new methods will address important challenges present in this and other inverse problems in cosmology, including scalability to large datasets, complexity of the forward model, and the need to quantify the uncertainty in the parameter estimation. On the other hand, this project will curate new datasets and make them publicly available to benchmark AI techniques for a new class of inverse problems, thus paving the way for future AI algorithmic innovation. Our project will bring together cosmologists, computer scientists, statisticians, and computational mathematicians, facilitating interdisciplinary and collaborative training.

The SkAI Institute is one of the National Artificial Intelligence Research Institutes funded by the U.S. National Science Foundation and Simons Foundation.
Information on National AI Institutes is available at aiinstitutes.org.