Improving star-galaxy separation with neural networks
Co-supervised Albert Sneppen, DAWN Summer Student (2015)
In this project a neural network has been designed to classify stars and galaxies, where an optimal use of information and structure of the neural network has been investigated. Thus, the project presents a computationally fast methodology with the freedom to choose between quantity and purity that can accurately predict more than 95 percent of stars and galaxies.
John R. Weaver
Postdoctoral Research Associate
My research interests lie almost exclusively within the realm of extragalactic astrophysics and cosmology. I use state-of-the-art optical and infrared observatories and surveys to study the lives of galaxies, and how their properties change over cosmic time. This includes detailed case studies of individual galaxies, as well as statistical analyses of large survey catalogs.