Astronomy Tea Talk
Speaker 1: Dr. Devontae Baxter
Title: Simulation-Driven Insights into the Impact of Incompleteness on Identifying and Interpreting Galaxy Protocluster Populations
Abstract: Galaxy clusters, the most massive gravitationally bound structures in the universe, represent the extreme end of hierarchical structure formation. While nearby clusters have been studied in great detail, the earliest stages of cluster formation — the protocluster phase — remain less well understood, largely due to a lack of large, spectroscopically-confirmed protocluster samples identified using uniformly selected apertures and overdensity tracers. Next-generation observatories, with their wide fields of view, ultradeep imaging, and high-resolution spectroscopy, will help address these challenges. However, cosmological simulations of galaxy formation remain essential tools for interpreting existing protocluster samples, making testable predictions, and guiding future protocluster surveys. In this talk, I will share insights from my recent work using high-resolution zoom-in simulations of galaxy clusters to quantify how biases in protocluster selection functions — typically favoring the most massive and star-forming galaxies — impact our ability to identify protoclusters and influence interpretations of these structures as sites of accelerated galaxy evolution in the early universe.
Speaker 2: Minjie Lei
Title: Morphology-based HI phase separation with scattering transform techniques
Abstract: Resolving the multi-phase structure of the diffuse interstellar medium (ISM) as traced by neutral hydrogen (HI) is essential to understanding the lifecycle of the Milky Way. However, HI phase separation is a challenging and under-constrained problem due to the limited availability of background continuum sources to measure HI absorption. In this talk, I will present a new statistical phase separation method that utilizes scattering transform (ST) statistics to encode HI spatial morphology structures into a compact and interpretable representation. This combined with a variational autoencoder (VAE) architecture allows us to learn the separation of HI phases directly from HI emission morphology information alone. Our result illustrates a clear physical connection between the HI morphology and HI phase structure, and unlocks a new dimension to improving future phase separation techniques by making use of both spectral and spatial information to decompose HI in 3D position-position-velocity (PPV) space.