Cosmic structure formation is hierarchical in nature, which means dark matter halos and the galaxies within them are assembled through successive accretion of smaller objects. Galaxy mergers are predicted to have a profound influence on many aspects of how galaxies form and evolve (e.g., assembling stellar mass, morphological transformation, triggering of AGN). However, quantifying the impact of merging has been problematic. One main difficulty is that it is very challenging to identify genuine mergers as traditional methods suffer from poor reliability and incompleteness.
In a number of publications this year, we have developed a novel merger detection method based on deep learning using convolutional neural networks. We applied this method to high quality imaging data sets from both real observations as well as state-of-the-art cosmological hydrodynamic simulations. Through this approach we have gained insights on the incompleteness and bias of visually identified merger samples from real observations. On the other hand, we have also learned valuable lessons on the degree of realism in simulated galaxies.
I will focus on the influence of mergers on the position of galaxies in the star-formation rate vs stellar mass diagram (the so-called galaxy star formation main sequence) across cosmic time and the connection between mergers and AGN which are still controversial issues in galaxy evolution. I will also discuss the upcoming ESA mission EUCLID in the context of galaxy merger studies.
| Attachment | Size |
|---|---|
 Shanghai_conference_LingyuWang.pptx | 7.05 MB |