A Deep Learning Method for AGILE-GRID Gamma-Ray Burst Detection

• Published in: The Astrophysical journal Vol. 914; no. 1; pp. 67 - 78
• Main Authors: Parmiggiani, N., Bulgarelli, A., Fioretti, V., Di Piano, A., Giuliani, A., Longo, F., Verrecchia, F., Tavani, M., Beneventano, D., Macaluso, A.
• Format: Journal Article
• Language: English
• Published: Philadelphia The American Astronomical Society 01.06.2021; IOP Publishing
• Subjects: Artificial neural networks; Astrophysics; Convolutional neural networks; Data analysis; Data search; Datasets; Emission analysis; Evaluation; Gamma ray bursts; Gamma rays; Gamma-ray astronomy; Gravitational waves; Neural networks; Real time
• ISSN: 0004-637X; 1538-4357
• DOI: 10.3847/1538-4357/abfa15

Abstract The follow-up of external science alerts received from gamma-ray burst (GRB) and gravitational wave detectors is one of the AGILE Team’s current major activities. The AGILE team developed an automated real-time analysis pipeline to analyze AGILE Gamma-Ray Imaging Detector (GRID) data to detect possible counterparts in the energy range 0.1–10 GeV. This work presents a new approach for detecting GRBs using a convolutional neural network (CNN) to classify the AGILE-GRID intensity maps by improving the GRB detection capability over the Li & Ma method, currently used by the AGILE team. The CNN is trained with large simulated data sets of intensity maps. The AGILE complex observing pattern due to the so-called “spinning mode” is studied to prepare data sets to test and evaluate the CNN. A GRB emission model is defined from the second Fermi-LAT GRB catalog and convoluted with the AGILE observing pattern. Different p -value distributions are calculated, evaluating, using the CNN, millions of background-only maps simulated by varying the background level. The CNN is then used on real data to analyze the AGILE-GRID data archive, searching for GRB detections using the trigger time and position taken from the Swift-BAT, Fermi-GBM, and Fermi-LAT GRB catalogs. From these catalogs, the CNN detects 21 GRBs with a significance of ≥3 σ , while the Li & Ma method detects only two GRBs. The results shown in this work demonstrate that the CNN is more effective in detecting GRBs than the Li & Ma method in this context and can be implemented into the AGILE-GRID real-time analysis pipeline.