Events Daily

Tuesday, October 22, 2024
      

Machine Learning Chalk Talk
Mia Morrell
Event Type: Special Seminar
Time: 11:00 AM - 12:15 PM
Location: Room 871
Abstract: This week, Mia Morrell will be presenting "From the Hopfield model and Boltzmann machines to maximum entropy models: attempts to understand real neural networks". Note the special location: room 871! MLCT is a weekly seminar hosted by the Physics department where anyone—students, postdocs, faculty—can present a 45-minute-ish lecture or discussion on machine learning or statistical inference. The setting is informal, chalkboard-focused (with optional slides), and interruptions encouraged. Topics can range from rough ideas to finished research, with the assumption that the audience has at least some working knowledge of ML. To see a weekly schedule of seminars, see https://dwh.gg/MLCT

cogsworth: A Gala of COSMIC proportions combining binary stellar evolution and galactic dynamics
Tom Wagg, University of Washington
Event Type: Informal Astro Talk
Time: 2:00 PM - 2:30 PM
Location: 726 Broadway, 940, CCPP Seminar
Abstract: Feedback from massive stars shapes the formation and evolution of galaxies. The majority of these massive stars are found in binaries, yet many parameters within binary stellar evolution remain poorly constrained. One avenue for improving constraints on these parameters is by using positions and kinematics of massive stars. Binary interactions can leave significant imprints on these parameters, ejecting massive stars rapidly from their birth sites. I will present a new code, cogsworth, which combines population synthesis and galactic dynamics self-consistently, providing the theoretical infrastructure necessary to make detailed predictions for the positions and kinematics of many different massive stellar populations. I will demonstrate how one can use cogsworth to make predictions for a range of observable binary products, from massive runaway stars and X-ray binaries, to supernovae and short gamma-ray bursts. I will show how cogsworth enables you to plot detailed evolution and orbits of specific binaries, track present-day positions of specific subpopulations and convert intrinsic populations to observables in Gaia. I also will share recent predictions I’ve made with cogsworth on how binary interactions can delay and displace supernova feedback in galaxies. This can reduce the efficiency of feedback close to star-forming regions, and potentially initiate star formation in new areas of the galaxy.