Projects
Concrete research programs in stellar spectroscopy, Galactic archaeology, Milky Way dynamics, and machine-learning methodology.
Applies neural networks to estimate stellar parameters and elemental abundances for 1.2 million giants from low-resolution LAMOST DR8 spectra.
The resulting value-added catalogue supports studies of stellar populations and the Milky Way's chemical evolution.
Develops a deep-learning framework to identify ex-situ stars using 6D kinematics and actions from Gaia DR3.
The project includes data preparation, classifier development, and analysis of the spatial distributions and relative contributions of identified stellar components.
Investigates a rotating stellar component in the bulge and halo using classification applied to Gaia DR3.
Test-particle simulations provide a framework for interpreting angular-momentum transfer from a decelerating Galactic bar.
In preparation
ANCHOR is an in-preparation framework for representation learning from stellar chemical information.
A fuller public description will be added when the methodology and results are ready.