I'm experienced in a variety of machine learning applications (including supervised and unsupervised) for galaxy morphological classifications and strong lensing detection. Meanwhile, I am also interested in and building up the abilities of ML applications to other astronomical studies such as data-driven simulations/emulations in imaging and spectroscopic data, lensed-quasar detection, etc. Astronomically, my research interests include galaxy evolution and formation, dark matter distribution in galaxies, correlation between SMBH and host galaxy, strong lensing and weak lensing.
I am currently working on several projects including:
physical properties extraction from spectroscopic data using ML
spectroscopic redshift estimation using ML
ML galaxy image emulation
machine-defined galaxy morphological classification systems
During my PhD study, I complete a comparison between several common supervised machine learning methods (e.g. k-nearest neighbour, support vector machine, random forest, neural networks, convolutional neural networks, etc) using DES imaging data as input on the morphological classification of galaxies (Cheng et al. 2020a). As the extension of my first project, I am building a galaxy morphology classification catalogue for DES Y3 data (Cheng et al. in Prep.). Meanwhile, I proposed an unsupervised machine learning technique on the detection of galaxy-galaxy strong gravitational lensing (Cheng et al. 2020b). This technique is then improved and used to explore the galaxy morphology using imaging data from the Sloan Digital Sky Survey (SDSS) (Cheng et al. 2020).
At National Tsing Hua University, I worked with Ing-Grey Jiang on the dark matter distribution of early-type galaxies (ETGs) (MSc Thesis). In our study, we assumed a functional total mass-to-light ratio and the numerical solution of a three dimensional Sersic profile in the Jeans equation, and concluded a discovery of 8 ETGs with an insignificant dark matter halo using data from Hubble Space Telescope at 555nm.