Conceitos Básicos
NuGraph2 is a powerful Graph Neural Network designed for low-level reconstruction of simulated neutrino interactions in a Liquid Argon Time Projection Chamber detector, offering high efficiency in background filtering and semantic labeling.
Resumo
NuGraph2 is a cutting-edge Graph Neural Network tailored for reconstructing simulated neutrino interactions in a Liquid Argon Time Projection Chamber (LArTPC) detector. It efficiently filters out background hits with 98.0% efficiency and accurately labels hits based on particle type with 94.9% efficiency. The network operates directly on detector observables without the need for arbitrary transformations or downsizing, providing flexibility across different detector technologies. By utilizing a multi-head attention message-passing mechanism, NuGraph2 ensures consistency between 2D representations while encouraging broader applications beyond the described tasks. The model's inference time is impressively fast at 0.12 s/event on CPU and 0.005s/event batched on GPU, making it suitable for real-time applications.
Estatísticas
Model inference takes 0.12 s/event on a CPU and 0.005s/event batched on a GPU.
NuGraph2 achieves 98.0% efficiency in background filtering and 94.9% efficiency in semantic labeling.