Abstract Details

Name: Abraham Mathews
Affiliation: Indian Institute of Space Science and Technology
Conference ID : ASI2024_716
Title : Exploring Alien Megastructures: Analyzing Transit Light Curves for Anomalies with Machine Learning
Authors : Abraham Mathews1, Vikram Khaire1,2, John Hoang3, Vishal Gajjar3,4
Authors Affiliation: 1 Indian Institute of Space Science & Technology, Thiruvananthapuram, Kerala 695547, India 2 University of California Santa Barbara, Santa Barbara, CA 93106-9530, USA 3 University of California, Berkeley, CA 94720, USA 4 SETI Institute, Mountain View, CA 94043, USA
Mode of Presentation: Poster
Abstract Category : Sun, Solar System, Exoplanets, and Astrobiology
Abstract : Within the framework of our assumptions concerning advanced extraterrestrial civilizations, we explore the possibility of colossal artificial structures designed for transportation or harnessing energy from host stars. These hypothetical megastructures, if prevalent in the galaxy, might leave discernible signatures within transit light curves. We employ Convolutional Neural Network-based machine learning (ML) techniques to predict the shape of transiting objects based solely on their respective transit light curves. A substantial deviation from a circular shape in the predicted outline could indicate the presence of potential alien megastructures. For training our ML algorithm, we created a library of 10^5 transiting light curves generated using different artificial closed shapes orbiting stars. Our algorithm provides excellent results with perfect light curves, and we get reasonably good predictions even for transit light curves forward modeled from the Kepler satellite with SNR > 50. Our next steps involve applying these trained algorithms to Kepler light curves to generate a catalog of potential transiting megastructure candidates. These candidates will be further followed up for anomalous and targeted radio signals through collaborative efforts with Berkeley SETI labs in search of extraterrestrial intelligence (SETI).