Machine Learning and High-Performance Computing for LArTPC

Gürkan Karaman; University of Iowa

Liquid Argon Time Projection Chamber (LArTPC) has been used in many neutrino experiments and is the technology of choice for the Deep Underground Neutrino Experiment (DUNE). Convolutional Neural Network (CNN) algorithms are becoming performant and prevalent choice to perform the classification of energy deposits and reconstructed particles as track-like or arising from electromagnetic cascades. The need to process billions of neutrino events with many machine learning algorithms creates computing challenge. In this talk, we will investigate Graphical Processor Units(GPUs) for accelerating time-consuming computing tasks.