|Time:||星期三, 12月 25, 2019, 10:30am|
|Title:||The Universe Machine: a brief introduction|
|Speaker:||Mr. Yunchong Wang (Stanford Univ.)|
In this talk we briefly introduce the methods and results of the Universe Machine. The Universe Machine is an empirical model that constrains galaxy-halo connection based on Bayesian inference of the halo star formation rate (SFR). It models the mapping of halo SFR from halo properties including halo mass (maximum circular velocity), assembly history, and redshift by constraining model parameters from observed stellar mass functions, cosmic SFR, UV luminosity functions etc. Compared to traditional abundance matching models, the Universe Machine does not impose any a priori assumptions about galaxy-halo connection, and it is also more flexible at exploring a broader range of physical processes than hydrodynamical simulations. The best-fit model of Universe Machine manifests crucial information about galaxy-halo connections, i.e. halo mass strongly influences star formation in star forming galaxies, combined with the time evolution of quenching towards lower mass halos, leads to the highest star formation efficiency in 10^12 Msun halos. The evolution of galaxy quenched fractions cannot be determined by single parameters such as halo virial temperature and cooling timescale, and reflects the more complex embedded feedback physics in galaxy formation. Central, quenched galaxies assemble their stellar population earlier than their satellite or star forming counterparts, in qualitative agreement with observations. Unequivocal detection of the correlation between halo SFR and halo mass accretion rate at all mass scales further justifies galaxy-halo connection. Extrapolation of the current best-fit model obtained from high mass constraints down to extremely low halo mass is a prospecting direction to study the universality of galaxy-halo connection.