Day 1350

Tuesday.

1712.03255
Painting galaxies into dark matter haloes using machine learning
Agarwal, Davé, Bassett

Use machine learning (ML0 to populate large DM-only sims with baryonic galaxies.  The ML framework takes input halo properties including halo mass, environment, spin, and recent growth history, and outputs central galaxy and overall halo baryonic properties including stellar mass, SFR, metallicity, and neutral hydrogen mass.  Apply this to the MUFASA cosmo hydro sim, and show that it recovers the mean trends of output quantities with halo mass highly accurately, including following the shape drop in SFR and gas in quenched massive galaxies.  However, the scatter around the mean relations is under-predicted.  Examining galaxies individually, at z=0 the stellar mass and metallicity are accurately recovered (sigma ~< 0.2 dex), but SFR and HI show larger scatter (sigma >~0.3 dex); these values improved somewhat at z=1,2.  ML quantitatively recovers second parameter trends in galaxy properties, e.g. that galaxies with higher gas content and lower metallicity have higher SFR at a given M*.  Testing various ML algorithms, find that none performs significantly better than the others.  Ensembling the algorithms does not fare better, likely because of correlations between the algorithms and the fact that none of the algorithms predict the large observed scatter around the mean properties.  For the random forest, find that halo mass and nearby (~200 kpc) environment are the most important predictive variables followed by growth history.  Find that halo spin and ~Mpc scale environment are not.  Finally, study the impact of additional inputting key baryonic properties M*, SFR, and Z, as would be available e.g. from an equilibrium model, and show that particularly providing the SFR enables HI to be recovered substantially more accurately.


1712.03644
Galaxy and Mass Assembly (GAMA): Blue spheroids within 87 Mpc
Mahajan, et al

Sample of 428 galaxies of various morphologies in 0.002<z<0.02 (8-87 Mpc) with panchromatic data from GAMA.  Find that BSph galaxies are structurally very similar to their passively-evolving red counterparts, but their SF and other properties such as color, age and metallicity are more like SF spirals than spheroids.  Show that BSph galaxies are statistically distinguishable from other spheroids as well as spirals in the multi-dimensional space mapped by luminosity-weighted age, metallicity, dust mass and sSFR.  Use HI data to reveal that some of the BSphs are further developing their disks, hence their blue colors.  They may eventually become spiral galaxies --- if sufficient gas accretion occurs --- or more likely fade into low-mass red galaxies.