|
My research connects stars, dust, and gas, with the goal of building a unified multidimensional picture of the Milky Way and nearby galaxies. I study the lifecycle of matter in galaxies: how gas and dust form molecular clouds, how stars are born from these environments, and how stellar evolution and feedback return material back to the interstellar medium. A central challenge is that galaxies and their internal structure, including molecular clouds, are complex, hierarchical, and only partially observed: we often infer intrinsically 3D processes from 2D projections, with incomplete distance and velocity information. To address this, I develop data-driven machine learning methods that combine heterogeneous observations into physically interpretable maps. These include Gaussian-process based methods for 3D and 6D astrocartography, which reconstruct the spatial and kinematic structure of interstellar material, and Gaussian mixture models with spatial Gaussian-process priors, which allow us to infer structured stellar and dust populations while preserving spatial coherence.
I will present my research built around these ideas, from the discovery and characterisation of the nearest molecular cloud to earth, Eos, to 3D dust maps of Milky Way dust and molecular cloud structure, and ongoing work on mapping the dust density in M33 using mixture models with GP priors. I will also discuss efforts to identify and characterise feedback-driven bubbles in both 2D observations and 3D data, linking statistical structure to the physics of star formation and feedback. Looking ahead, I will outline how these tools can support future observational programs, including the NASA SMEX mission, Eos. Together, this work seeks to move from static snapshots of the Milky Way and nearby galaxies toward a unified, multidimensional picture of how stars, dust, and gas interact across the lifecycle of matter in galaxies and how they evolve.
|
|