Research
Our key research areas are described below
Cosmology, Dark Ages and Cosmic Dawn
The universe started from a big bang almost 14 billion years ago. After around a million year, the universe consisted of mainly neutral hydrogen gas. Because the gas did not emit much light, the period after around a million year is called the dark ages. The gas was almost homogeneous with small overdensities that gave rise to galaxies during the cosmic dawn. The stars from these galaxies ionized almost all intergalactic hydrogen during the epoch of reionization which ended around a billion year after the big bang.
We study these early phases of the universe using theoretical modeling, simulation and cosmological observations. For example, we use 21cmFAST to simulate the neutral hydrogen and work with the data of the LOFAR telescope to understand some systematic effects in the observations of the epoch of reionization.
Galaxies, Galaxy Clusters and Dark Matter
Galaxy clusters are the largest gravitationally bound structures in the universe. A cluster contains thousands of galaxies, intergalactic gas constituting the intra-cluster medium (ICM) and dark matter. Dark matter accounts for almost 80% of the total mass of a cluster while the hot gas in the ICM accounts for around 15% and the remaining few percent is accounted for by all the galaxies and their stars. Galaxies and galaxy clusters are thus very important probes of dark matter and large-scaled plasma.
We study radio and x-ray observations of galaxy clusters. For example, we use radio observations of the MeerKAT telescope and x-ray observations of the Chandra telescope to study minihalos, small radio-emitting regions in the centers of galaxy clusters.
Radio Astronomy Antennas
Radio telescopes and arrays are one of the most important probes for learning about the structure and evolution of the universe and its largest components, the galaxies and their clusters. In order to extract scientific information from radio observations, it is essential to know the telescope precisely. Until recently, it was enough to know the direction (toward sky) independent systematic effects of a radio telescope, but we must understand the direction dependent effects accurately for the modern wide-field cosmological observations.
We try to understand the direction dependent systematic effects of radio telescopes to facilitate astronomical observations. For example, we work with the observations of LOFAR and MeerKAT to understand the effects of their primary beam.
Machine Learning in Astronomy
The whole universe is the laboratory of an astronomer and astrophysicist. All we need are telescopes and supercomputers to store their observations. But the amount of data is so large that it is humanly impossible to understand many of the modern scientific problems scientists try to solve. Therefore, machine learning is very important for classifying, modeling and finding clusters and groups in astronomical data. Almost every field of modern astronomy has been touched by the recent breakthroughs in artificial intelligence.
We study various applications of machine learning to astronomical data. For example, we use semi-supervised deep-learning models for classifying radio galaxies imaged by the VLA telescope as part of its FIRST survey.