COSMOS2020: Manifold learning to estimate physical parameters in large galaxy surveys

We present a novel method for estimating galaxy physical properties from spectral energy distributions (SEDs) as an alternative to template fitting techniques and based on self-organizing maps (SOMs) learn the high-dimensional manifold of photometric catalog. The has previously been tested with hydrodynamical simulations in Davidzon et al. (2019, MNRAS, 489, 4817), however, here it is applied real data first time. It crucial its implementation build SOM high-quality panchromatic set, thus we selected “COSMOS2020” catalog this purpose. After training calibration steps COSMOS2020, other galaxies can be processed through SOMs obtain estimate their stellar mass star formation rate (SFR). Both quantities resulted good agreement independent measurements derived more extended baseline and, addition, combination (i.e., SFR vs. diagram) shows main sequence star-forming that consistent findings previous studies. discuss advantages compared traditional SED fitting, highlighting impact replacing usual synthetic templates collection empirical SEDs built by “data-driven” way. Such approach also allows, even extremely large sets, efficient visual inspection identify errors or peculiar types. While considering computational speed new estimator, argue will play valuable role analysis oncoming large-area surveys such Euclid Legacy Survey Space Time at Vera C. Rubin Telescope.