Past Event: Babuška Forum

Harnessing Machine Learning to Study How Stars Form

Professor Stella Offner, Associate Professor, Astronomy, Oden Institute, UT Austin

Abstract

Forming stars interact with their birth environment via radiation and mass ejection, including through stellar winds and outflows. This "stellar feedback" impacts the efficiency at which stars form and the resulting distribution of stellar masses. Historically, signatures of stellar feedback were identified "by eye.” However, this approach is challenging, time-consuming and subjective. In this talk, I will show how a combination of state-of-the-art numerical simulations together with machine learning can be harnessed to identify features in telescope observations and to predict underlying physical properties. I will present results obtained using a convolutional neural network method, 3D Convolutional Approach to Structure Identification (CASI-3D) and a Denoising Diffusion Probabilistic Model.  Meanwhile, unsupervised machine learning approaches are powerful tools for identifying patterns in high-dimensional data. I will briefly discuss results using unsupervised methods to study the evolution of star-forming gas.

Biography

Stella Offner is an Associate Professor of Astronomy at UT Austin. She received her PhD in Physics from UC Berkeley. She was a NSF Astronomy & Astrophysics Prize Postdoctoral Fellow at the Harvard-Smithsonian Center for Astrophysics and a NASA Hubble Postdoctoral Fellow at Yale.