SLAC researchers develop SPINE machine-learning tool for neutrino detection

Researchers at the Department of Energy’s SLAC National Accelerator Laboratory have developed SPINE (Scalable Particle Imaging with Neural Embeddings), a machine-learning program designed to reconstruct neutrino interactions from complex detector images.

Neutrinos are nearly massless, elusive particles that interact weakly with ordinary matter. While trillions pass through the Earth every second, capturing and interpreting their interactions is difficult due to overlapping signals and background activity. Traditional algorithms have struggled to process the complicated data produced by large-scale detectors.

SPINE is currently being deployed in several major neutrino experiments, including the Short-Baseline Neutrino program and the Deep Underground Neutrino Experiment (DUNE). The tool aims to help scientists investigate fundamental physics questions, such as the nature of sterile neutrinos, neutrino masses, the properties of dark matter, and the imbalance between matter and antimatter in the universe.