Machine learning black hole

Measuring black hole mass is critical for many areas of astrophysical research from cosmology to galaxy evolution. We know about half a million black holes with a mass measured by traditional approaches using spectroscopic observations, representing decades of state-of-the-art community efforts. However, LSST will discover ~a billion new black holes, for the majority of which spectroscopic observations will not be practical. We are building deep learning algorithms to seek out an alternative way for estimating black hole mass using only the accretion disk time series. Check out our publications.