Rapid Gamma-Ray Burst Localization aboard the e-Astrogam Satellite using a 3D Convolutional Neural Network

Ruoxi Shang, Andreas Zoglauer

BayLearn: Bay Area Machine Learning Symposium
October 16, 2019

Abstract: The discovery of the association of a short gamma-ray burst (GRB) with gravitational waves from the neutron-star-neutron-star merger GW170817 gave rise to the requirement to localize GRBs in close to real time even with the most sensitive but also data-analysis-wise most complex gamma-ray space telescopes, the Compton telescopes. Here we report on the implementation and testing of a 3D convolution neural network trained to localize the origin of Compton-scattered gamma-rays from GRB’s on the sky. The ultimate goal of this project is the implementation of a neural network for GRB localization in a FPGA aboard a satellite such as the envisioned e-Astrogam.

This poster was presented as part of the undergraduate internship research project Enabling future gamma-ray space missions, offered through BIDS and URAP.

Presented at the BayLearn Symposium (Bay Area Machine Learning Symposium) on October 16, 2019.



Featured Fellows

Andreas Zoglauer

Space Sciences Laboratory
BIDS Alum - DATA SCIENCE FELLOW