Title: From Planets to Exoplanets: Local Interactions, Population Clustering, and Planet Formation Pathways

Speaker: Dr. Yi Duann (National Central University)

Time: 10:00am, June 5, 2026

Location: 3-402, PMO Xianlin Campus

Abstract: This presentation explores how observable signatures of exoplanets can be connected to their underlying physical processes and formation histories. Starting from a case study of the ultra-hot Jupiter WASP-189b, we investigate whether observed near-ultraviolet anomalies can be explained by star–planet interactions through magnetohydrodynamic simulations. The results highlight the complexity and degeneracy in interpreting individual systems. To overcome these limitations, we move to a population-level approach using machine learning. By applying Gaussian Mixture Models to dynamical parameters, we identify intrinsic clustering structures in close-in exoplanets. These clusters are then mapped onto synthetic populations generated from pebble-accretion models, allowing us to link observed planetary properties to formation parameters such as gas accretion efficiency, formation timing, and core composition. We find that exoplanet populations are not randomly distributed but instead reflect distinct formation pathways. However, degeneracies remain, as similar observable properties can arise from different evolutionary histories. This motivates the need for a statistical framework that connects observations to formation physics. Finally, we outline ongoing and future work aimed at integrating atmospheric composition with population clustering, and developing machine-learning models capable of predicting formation pathways and internal properties such as the ice–rock ratio. This approach aims to bridge the gap between observation and theory, enabling a more predictive understanding of planet formation.