Deep-learning trigger system for high-energy neutrinos

Christian Glaser. Foto: Camilla Thulin.


Projekttitel: Deep-learning trigger system for high-energy neutrinos
Huvudsökande: Christian Glaser, avdelningen för högenergifysik
Beviljade medel: 600 000 kr över två år

Neutrinos are the perfect cosmic messengers. Ultra-high energy (UHE) neutrinos will provide insights into the inner processes of the most violent phenomena in our universe, those that happen in the vicinity of super-massive black holes (e.g., in active galactic nuclei), in neutron star mergers or gamma ray bursts. The detection of these ghostly, extremely energetic elementary particles would be one of the most important discoveries in astroparticle physics in the 21st century. To be able to measure the low flux of UHE neutrinos on Earth, a new detector technology has been developed, instrumenting polar ice sheets with radio antennas to search for neutrinos passing through the ice. The first detector of substantial size is currently being constructed in Greenland.

In this project we address the critical aspect of increasing the detector sensitivity by making the detector smarter. We will bring artificial intelligence (with dedicated low-power on-device computing) to the radio detector stations, which will increase their sensitivity to neutrinos substantially at negligible additional hardware costs. This will increase the number of detectable neutrinos and enable a more detailed study of the high-energy universe.

Senast uppdaterad: 2022-01-14