Speakers
Description
For more than two decades, the MAGIC telescopes continuously accumulate
significant amounts of data. However, the analysis of this data poses critical
problems due to its volume exceeding existing data curation capacities. This
criticality induces the demands for the utilization of AI methods to enhance and
accelerate the analysis process. Thus, MAGIC utilizes random forests for an ac-
celerated and robust reconstruction of the energy and direction of the measured
particles.
Consequently, efficient analysis performed with the respective AI methods re-
quires the development of a tool that ensures traceability as well as repro-
ducibility. Therefore, we present the database-driven tool autoMAGIC, capa-
ble of coordinating the use of random forests for large-scale datasets. Based
on the analysis specifications, autoMAGIC runs the respective tools for choos-
ing suitable training data, training and testing the random forest, and storing
the outputs for further processing over multi-year datasets. Furthermore, we
present long-term lightcurves performed with autoMAGIC, demonstrating the
use of autoMAGIC to acquire labor-intensive AI-based results efficiently.
