David Chemaly

University of Montreal

Despite being less massive by orders of magnitude, supermassive black holes (SMBH) have a profound impact on their host galaxy. This effect is depicted by a variety of relations, the most notable of them being the M-sigma relation which shows that the mass of a SMBH is directly proportional the mass of its host galaxy. Although we are aware of this co-relation, the science community has yet to identify its cause. One of the leading hypotheses is the co-evolution through cosmic time of the SMBH and its host galaxy.

Considering that light has a finite speed, the more distant an object is observed the further we are looking at it back in time. Therefore, measuring the mass of SMBH at high distances would allow us to prob the state of the Universe at a younger age and confirm or deny the existence of a co-evolution. Sadly, measuring the mass of a SMBH requires an extremely high spatial resolution that even our most state-of-the-art telescopes can’t achieve at these higher distances. Luckily, one way to solve this limitation is by observing SMBHs through gravitational lenses. At the price of greatly distorting the observed SMBH, these cosmic lenses will magnify it.

My work consists of training a Convolutional Neural Network (CNN) capable of measuring the mass of SMBHs at higher distances by utilising the magnification of gravitational lensing. Such a tool will allow us to better understand the evolution of SMBHs, galaxies and more broadly our Universe through cosmic time.

Example of a galaxy-galaxy strong cosmic lens. The luminous red galaxy (center of the image) has gravitationally lensed a much further blue galaxy (blue wrap). The alignment between these galaxies is so precise that the background galaxy got distorted into a horseshoe. Credit: ESA/Hubble & NASA.