Speaker
Description
Dwarf galaxies evolve in diverse environments and serve as important probes of dark matter physics and galaxy formation. In this talk, I will present the environmental evolution of dwarf galaxies within a two-component self-interacting dark matter (SIDM) model featuring mass segregation. In this model, energy exchange between heavy and light dark matter particles causes the heavier component to move inward, increasing inner halo densities, accelerating gravothermal evolution, and expanding galaxy sizes. This mechanism naturally addresses the unexpectedly high lensing efficiency of cluster substructures, the ultra-compact dark subhalos inferred from strong lensing, and the enhanced clustering of diffuse dwarf galaxies. To generate predictions across environments, we combine cosmological cold dark matter simulations with a parametric SIDM model. The former captures halo assembly and environmental dependence, while the latter describes SIDM as a controlled deformation of halo evolution paths. Integrating self-interaction effects along these paths enables efficient predictions for both isolated dwarf galaxies and cluster subhalos within a common framework. More generally, the path-deformation perspective motivates a probabilistic extension of the inference framework. We therefore formulate galaxy formation as a stochastic dynamical process and develop a path-measure framework based on hierarchical halo graphs and graph neural networks. This framework provides a natural setting in which environmental effects and SIDM physics can be encoded as controlled deformations of galaxy formation histories.