Home
- Seminars and Workshops
- Institute of Astronomy and Astrophysics
- Location
R1203 of the Astronomy-Mathematics Building, National Taiwan University
- Speaker Name
Sheng-Chieh Lin (University of Kentucky)
- State
Definitive
- Url
Estimating Cluster Mass from SDSS Broad-band Images by Deep Learning
2020-12-24 14:00 - 15:30
The cluster mass is one of the fundamental properties of galaxy clusters. In optical surveys, several measures can be taken to estimate the cluster mass, including, for example, utilizing the velocity dispersion of the cluster galaxies, or the weak gravitational lensing. On the other hand, deep learning provides an unique way to predict the cluster mass by approximating the underlying complicated mappings from inputs (images or catalogs) to outputs. In this talk, I will present a semi-supervised machine learning model which estimates the cluster mass directly from the SDSS ugriz-band images. I will also show the visualization of the model which can be used to interpret the network output.
