Estimating the cold molecular gas mass of strongly lensed z=2.9 galaxies with ALMA CO(3-2) data.
Applying a Bayesian hierarchical model and NUTS to 6D phase-space information from halo stars to constrain the mass distribution of the Milky Way's dark halo.
We estimate stellar ages from labels containing information from both stellar evolution (intrinsic to the star) and Galactic evolution (environment, extrinsic to the star). We develop a probabilistic machine learning estimator and use SHAP values to explain and quantify how much of the age information is coming from intrinsic sources.
We develop a feature attribution method can be applied to outlier detection methods that explains why a particular data point is an outlier. It works for both small-scale and large-scale features, and is more easily interpretable than existing methods.
Camera control software for the narrowband upgrade to the Dragonfly Telephoto Array.