Abstract Details

Name: Yogesh Wadadekar
Affiliation: NCRA-TIFR
Conference ID: ASI2020_560
Title : Some novel applications of deep learning to astronomical problems
Authors and Co-Authors : Yogesh Wadadekar (NCRA-TIFR) Omkar Bait (NCRA-TIFR) Shraddha Surana (ThoughtWorks) Harsh Grover (BITS, Pilani) Lijo T. George (NCRA-TIFR)
Abstract Type : Poster
Abstract Category : Instrumentation and Techniques
Abstract : Machine Learning algorithms based on neural networks are good tools for both classification and regression problems, in many data rich fields of science. In the present era, astronomy has become extremely data rich and machine learning can now be applied to many different problems in the astronomy domain. A specific machine learning architecture called deep learning with many hidden layers in the network has emerged as a very powerful and accurate option in many areas. Excellent implementations of deep learning techniques are also freely available. We have effectively applied supervised deep learning techniques to address a number of diverse problems which include (1) predicting star-formation histories of galaxies using flux measurements in 21 broad bands from UV to far-infrared (2) predicting the bulge-to-total luminosity ratio of galaxies from $gri$ color composite images and (3) classifying radio galaxy images into subclasses - compact, FR-I, FR-II, and bent tail. In all these applications, our deep learning models are able to make robust predictions with low error rates and in real time once the model has been trained. We highlight the challenges faced in terms of data size, availability, features, processing ability and importantly, the interpretability of results for these problems. In coming years, as the data tsunami in astronomy grows exponentially with the commissioning of facilities like the the SKA and the LSST, increased use of machine learning to understand the underlying physics in the information captured seems inevitable.