Dr. Madhurima Choudhury

The redshifted 21-cm line of neutral hydrogen (HI) promises to be an exceptionally sensitive probe for exploring various phases in the evolution of our Universe. The epoch of reionization (EoR) marks a pivotal transition in the high-redshift Universe, eventually leading to the complete ionization of neutral hydrogen in the intergalactic medium (IGM). Direct observations of the 21-cm line of HI offer a unique opportunity to map the early Universe across cosmic epochs, providing profound insights into the state of the IGM and the morphology of ionized structures sculpted by the first sources of light. Additionally, these observations allow for a deeper understanding of the origin and evolution of these first-generation sources. However, detecting this faint signal is challenging and is the primary science objective for numerous ongoing and upcoming low-frequency radio interferometers, including LOFAR, MWA, HERA, SKA, and global signal experiments like SARAS, EDGES, and REACH. In this talk, I will focus on the importance and application of machine learning (ML) techniques in deciphering the EoR from both theoretical and observational perspectives. I will elaborate on the implementation of neural networks to establish a framework capable of extracting and constraining IGM properties, such as the bubble size distribution and reionization histories, from 21-cm power spectrum measurements. Additionally, I will discuss the use of ML methods to extract reionization parameters from future observational images. Finally, I will highlight the potential uses of ML methods in the future, which will complement and enhance our understanding of the Universe.