2024

Strategic Projects

@Physics/Astronomy

#Self-supervised Learning

#Semi-supervised Learning

#Sensitivity enhancement

#Data Analysis based on AI

Project Summary

The aim of the ADAPT project was to develop a novel analysis strategy for the KM3NeT astroparticle physics experiment, using semi-supervised learning, to minimise the dependence on labelled data coming from Monte Carlo simulations. KM3NeT, a neutrino telescope located deep in the Mediterranean Sea, consists of a 3D array of Cherenkov light detectors that capture the passage of charged particles induced by neutrino interactions. A key challenge in KM3NeT data analysis is the substantial light background caused by K40 decays in seawater, which surrounds each neutrino-induced event. We will present our results on the implementation of this innovative analysis strategy.

 

Yvonne Becherini
yvonne.becherini@u-paris.fr

  • Professor at Université Paris Cité

Ankur Sharma
Pranju Goswami
Maximilian Eff
Enzo Oukacha

Projects in the same discipline