Hello There!

Own Picture Welcome to my digital home! I am Aritra (aw-ree-tro), a Ph.D. candidate in the Astronomy department at Yale University and I work in Prof. Meg Urry's group. Currently, I am applying the latest advances in convolutional neural networks in combination with traditional techniques to study the co-evolution of galaxies and black holes. Over time, I have worked on blazars, galaxy evolution, high energy physics and condensed matter physics -- To learn more about my Research, scroll up/down, open up the Menu! I am and have also always been heavily involved in science outreach and various student advocacy groups. Look at my CV to learn more!!

I was born and brought up in Kolkata, a laid-back old-school paradoxical mega-metropolis on the banks of the Ganga, in eastern India. I took up physics for my bachelors at Presidency University, Kolkata, a place to which I owe a lot both professionally and personally. After 3 years at Presidency, I moved to Groningen, a beautiful small university-town in northern Holland for my masters. After two years in Holland, I decided to cross the Atlantic and come to Yale to pursue my PhD in Astrophysics. When I am not (astro)physics-ing, you may find me cooking/baking, biking, driving around small towns on the East Coast or listening to music. The below icons should give you an indication of other things that I enjoy.

  • Education
  • Computing
  • Food
  • World Affairs
  • Travel
  • News
  • Music

Current Research

One of the most groundbreaking discoveries in modern astrophysics has been the fact that all galaxies above a mass of $\sim10^{10}M_{\odot}$ host a supermassive black hole at their center. These black holes primarily grow through the accretion of gas and during phases of extreme accretion, become visible as Active Galactic Nuclei (AGN); depositing large amounts of energy into the surrounding medium, often comparable to the binding energy of the galaxy. Although observational evidence has clearly identified that some AGN inject considerable energy/momentum into their host galaxies, the global impact of AGN "feedback" on the evolution of their host galaxies is not at all well established. For eg. one of the most popular ways to investigate the co-evolution of AGNs and their host galaxies is to assess the impact that AGNs have on star formation in these galaxies. A plethora of studies have investigated the star formation rates (SFR) of AGN hosts; however, these studies have had widely varying conclusions with claims of positive, negative and no correlation between AGN X-Ray luminosity ($L_X$) and SFR. These conflicting results can largely be attributed to low number statistics, use of a single FIR band to derive SFR thereby resulting in large uncertainties and selection biases due to different AGN selection techniques used in each study. Thus, an accurate understanding of why, how and when black holes affect the evolutionary pathways of their host galaxies remains one of the major outstanding questions in modern astrophysics.

Very large samples are crucial in order to conclusively understand AGN-galaxy co-evolution as they provide a `macro' view of both the AGN and host galaxy populations. Furthermore, a combination of different selection techniques and high-quality multi-wavelength data are essential to any attempt at trying to disentangle the large number of physical parameters such as mass, environment, morphology, and star formation rate. In order to overcome this hurdle, we are using a 'wedding-cake' sample of surveys of different areas and depths to properly sample AGN across the $L_X-z$ plane. To analyze this large dataset, we are using the latest advances in Machine Learning combined with traditional techniques. As a first step in this direction, we have developed Galaxy Morphology Network (GaMorNet), a Convolutional Neural Network (CNN) that does not need a large training set of real galaxies and has been sucessfully applied to both $\sim 100,000$ SDSS and $\sim 20,000$ CANDELS galaxies achieving net misclassification rates of $\lesssim 5\%$. We used the morphologies inferred by GaMorNet to study the quenching of star-formation in these galaxies. To learn more about GaMorNet, pelase use the links below.

Currently, I am working to process AGN through GaMorNet, in order to use morphological information to study the co-evolution of black-holes and galaxies using a large, comprehensive dataset.

Previous Projects I have been involved in

During my undergraduate years, I wanted to explore the various avenues available within physics and hence I worked on a diverse range of topics during the first three years of my college education. Slowly, I developed a liking for computation and extragalactic astrophysics. The list below lists my past work in reverse chronological order.

  • Diffusive Shock Acceleration

    For my masters thesis, I worked on incorporating Fermi Acceleration within CR Propa. CR Propa happens to be a public astrophysical simulation framework for cosmic rays. Being the primary mechanism for charged particles to gain non-thermal energy in astrophysical shock waves, fermi acceleration has widespread applications (including the shock physics of Supernova Remnants, which happen to be potential candidates for high energy neutrinos). You can find more details about this work here. ~~Publication~~
  • Blazar Light Curves

    In the summer of '15, I developed a python pipeline (which used Fermi Science Tools) to reduce Fermi data and studied the time-variability of quasars at GeV energies with this. Later, in the summer of '16, I used my code to reduce Fermi data for a related project led by an MSc student where we studied the symmetry property of blazar light curves in the gamma-ray and optical wavelengths in order to infer about the radiation mechanism(s) and pinpoint the location of the emission region. ~~Publication.~~

  • Belle II Feasibility Study

    The Belle II is an upgrade of Belle, an ep collider. I performed a feasibility study for Belle II while I was in Tokyo for 3 months in '14 as a participant in UTRIP.I studied tensor coupling in the leptonic decays of Tau-. I was able to show that present data from Belle and future data from Belle II should be able to improve upon the uncertainty in measuring the tensor interaction substantially

  • Fractals

    This one's close to my heart as this was my first research project! It all started when in '12, as a part of my summer research at IACS, I started reading up on fractals. Intrigued, I went ahead and wrote some C scripts to reproduce classic fractal sets like the Mandelbrot & Julia sets. Enthralled at having achieved this successfully, I went on to work on the aggregation mechanism of real life fractal nano-structures and continued this part-time for the next two years along with my other research & studies. ~~Publication~~

Publications & Other Material

I have a very common name and just searching by my name in ADS will not give you correct results. The best technique is to search using my ORCiD orcid:"0000-0002-2525-9647" or use the links below. The personal list that I maintain also has links to material that is not available via other public platforms.