A Novel Dynamical Modeling Method Based on the Data-driven Distribution Function

I will present a novel data-driven distribution function (DF) method, which constructs the DF of a tracer population empirically from observed 6D kinematics. It can maximize the usage of observation data but remove the dependence on the assumption of functional form or orbit library in conventional DF- or orbit-based methods. Its flexibility and efficiency make it a preferable alternative to the popular Jeans equation method, thus ensuring broad applications in the Gaia era. As the first application example, I applied the new method to measure the MW potential using satellites and globular clusters as tracers, where the limited sample size and large observational errors of tracers in the outer halo make this method particularly attractive for its efficiency.