Dr. Dalya Baron (Stanford): “Unifying the Digital Sky in the Era of Survey Astronomy”
Speaker: Dr. Dalya Baron (Stanford)
Title: “Unifying the Digital Sky in the Era of Survey Astronomy”
Video:
Abstract: Throughout history, new ways of observing the night sky have led to discoveries that reshaped our understanding of the Universe. Today, the next such shift is being driven not by a single instrument, but by an ecosystem of large-scale imaging and spectroscopic surveys observing billions of astronomical objects across wavelengths and time, exemplified by facilities like the Vera C. Rubin Observatory. In this regime, discovery requires integrating observations across heterogeneous imaging systems, spatial scales, and temporal cadences. In this talk, I will frame astronomy as an extreme imaging and inference problem: fusing data with varying resolutions, noise properties, point-spread functions, and spectral coverage. I will describe how the astronomical community is leveraging data science and AI to meet these challenges, supporting tasks from cross-survey data fusion to the automated discovery of new trends, classes, and outliers. I will finish by describing a new project within the Center for Decoding the Universe at Stanford that aims to build a single data-driven model to unify observations from ultraviolet to radio wavelengths of nearby galaxies, providing us with the most coherent picture of their stars, gas, and molecules.
Bio: Dalya is a research scientist at Stanford’s Kavli Institute for Particle Astrophysics and Cosmology (KIPAC). She is also part of the Center for Decoding the Universe within Stanford Data Science, which aims to unlock the physics of the Universe by developing innovative approaches for extracting insight from vast, multimodal datasets. As an observational astronomer, she uses some of the world’s largest ground- and space-based telescopes to study galaxies and their constituents: stars, gas, dust, and supermassive black holes. She is particularly interested in mining large astronomical datasets to uncover new trends and previously unknown classes of objects.
