01/22/2025
By Lynne Schaufenbil
Abstract: The next wave of astronomical data is set to arrive from new observatories like the Vera C. Rubin Observatory and NASA's Roman Space Telescope. Traditional methods for handling this data are becoming outdated due to the sheer amount and complexity of the information these facilities will collect. Artificial Intelligence (AI) offers a promising solution to manage and analyze these large datasets. However, integrating AI into astronomy isn't straightforward. Current AI techniques, developed mainly for simpler and more structured data, must be retooled to handle the unique challenges presented by astronomical data. This talk will focus on the need of developing AI tools tailored for astronomy and will discuss the collaborative efforts required to create robust, domain-specific AI applications that produce scientifically accurate results.
About: Cecilia Garraffo is the founding director of AstroAI, a cutting-edge research institute dedicated to advancing astrophysics through the application of artificial intelligence at the CfA. Originally from Argentina, she obtained her MS in Astronomy from La Plata National University in 2005, delving into gravitational field theories and various extensions of General Relativity during her graduate studies. In 2010, she earned her Ph.D. in Physics from the University of Buenos Aires, solidifying her expertise in theoretical physics and cosmology.
After completing her doctorate, Cecilia became a research associate at Brandeis University, where she continued her research in theoretical physics and cosmology. She joined the Center for Astrophysics | Harvard-Smithsonian in 2013 as a postdoctoral fellow in stellar astrophysics. During her postdoc, she employed cutting-edge machine learning algorithms to conduct data-driven astrophysical research, with a specific focus on stellar evolution, activity, X-ray emission, rotation, and the intriguing phenomenon of star-planet interaction. In 2018, Cecilia became part of the Institute for Applied Computational Science at Harvard University, where she formalized her skills in deep learning for astrophysical problems and taught the Master's Program in Data Science. In this period she directed her attention towards uncertainty quantification in deep neural networks, focusing on the development of probabilistic models and Bayesian neural networks to extract fundamental stellar properties from observable quantities. PhD in Physics, MS in Astronomy
For the Zoom link, please contact Lynne_Schaufenbil@uml.edu. Thank you!