Exhaustive Neural Importance Sampling applied to Monte Carlo event generation
This addresses a computational bottleneck in neutrino physics by making Monte Carlo event generation faster and more practical for researchers in that field, though it appears incremental as it builds on existing rejection sampling techniques.
The authors tackled the slow and impractical Monte Carlo generation of neutrino-nucleus cross-section models by introducing Exhaustive Neural Importance Sampling (ENIS), a method using normalizing flows to automatically find proposal densities for rejection sampling, resulting in improved efficiency for neutrino oscillation experiments.
The generation of accurate neutrino-nucleus cross-section models needed for neutrino oscillation experiments require simultaneously the description of many degrees of freedom and precise calculations to model nuclear responses. The detailed calculation of complete models makes the Monte Carlo generators slow and impractical. We present Exhaustive Neural Importance Sampling (ENIS), a method based on normalizing flows to find a suitable proposal density for rejection sampling automatically and efficiently, and discuss how this technique solves common issues of the rejection algorithm.