Extracting information from stochastic fields is a ubiquitous task in science. However, from cosmology to biology, it tends to be done either through correlation analyses, which is often too limited to capture the full complexity, or through the use of neural nets, which require large training sets and lack interpretability. I will present a new approach that borrows ideas from both extremes and stands nicely in between. I will use various examples in astrophysics and beyond, including its recent application to weak lensing data, to demonstrate its efficiency, interpretability, and advantages over traditional statistics.
BIO
Sihao Cheng is a Member of the Institute for Advanced Study and a visiting fellow of the Perimeter Institute. He obtained Ph.D. from Johns Hopkins University and Bachelor degree from Peking University. He uses innovative ideas to analyze survey data. His interest spans from cosmology to (exo)planets.
