Events Daily

Tuesday, November 19, 2024
      

Machine Learning Chalk Talk
Clark Miyamoto
Event Type: Special Seminar
Time: 11:00 AM - 12:15 PM
Location: 726 Broadway, 940, CCPP Seminar
Abstract: This week, Clark Miyamoto will be talking about "Two perspectives on how renormalization shows in ML". To see a weekly schedule of seminars, see https://dwh.gg/MLCT

Statistical Models of the Spatial, Kinematic, and Chemical Complexity of Dust
Andrew Saydjari, Princeton University
Event Type: Astro Seminar
Time: 2:00 PM - 3:15 PM
Location: 726 Broadway, 940, CCPP Seminar
Abstract: Dust pervades the Galaxy, impacting all observations and filling the space between the stars. Given its ubiquity, the density, motion, and chemistry of dust traces structure formation and destruction across a multitude of scales. In this talk, I will present a new 3D dust map reaching out to 10 kpc that represents an order of magnitude increase in angular resolution (1 arcmin) and discuss the methodology and photometry underpinning this result. Turning to spectroscopy, I will show that our new catalogs of diffuse interstellar bands (DIBs) in APOGEE spectra serve as precision kinematic tracers, that also have distance upper limits, which is crucial for understanding the dynamics of the Milky Way. Finally, I will show these DIBs provide a new window into the chemical diversity and variation of Galactic dust. Throughout, I will emphasize the key role of (linear) Bayesian component separation methods and data-driven priors.