Title: Opportunities for Applications of Deep Learning in Cosmology
Abstract:
A primary goal of modern cosmology is to map complex large-scale observations to simple theories. This contrast of scale between theory and data inevitably necessitates a computational approach. This may involve a compressive analysis of observational data, massive simulations, or a search for rare events. In this talk, I will argue that recent advances in machine learning and in particular deep learning can significantly change the current practice in “all” of these fronts. I will review several of our past and ongoing collaborations and identify exciting opportunities for interdisciplinary research.
TSI Seminars take place weekly during the Fall and Winter terms. TSI seminars are intended to be accessible to scientists from the entire breadth of backgrounds at TSI, including, Physics, Planetary Science, Geology, Atmospheric Science, and Astrobiology. Our seminar series is partially funded by the Centre de recherche en astrophysique du Québec (CRAQ).