Fabula’s Geometric Deep Learning helping detect neutrinos
Fabula’s patented technology, Geometric Deep Learning is already showing best-in-class success rates for detecting fake news. This innovative technology has wider uses than just fake news detection - in this case it’s helping detect neutrinos. Federico Monti, in collaboration with colleagues from New York University, Berkeley and Imperial College, worked with the Lawrence Berkley National Laboratory on data acquired by the IceCube Neutrino Observatory at the South Pole.
The research work carried out by Fabula’s Federico Monti, a member of Prof. Michael Bronstein group, earned him the award for the best scientific contribution assigned by ICMLA, the most important international conference in the field.
According to Federico Monti: “It is extremely difficult to effectively process the data collected by the IceCube detectors with traditional Deep Learning Algorithms (which are useful to process data defined on regular grids such as images, videos or audio messagges). The main obstacle lies in the irregularity of the network of sensors positioned in the observatory, presenting different levels of density in different regions of the ice. The new GCNN algorithms are ideal for this purpose, as they are designed to address irregular structures such as graphs or surfaces”.
For more details, please see the press release from Università della Svizzera italiana - https://www.usi.ch/en/feeds/9771.