This project is developed to provide classification solution for Background Alpha and Neutrino Beta Events from the Jinping Neutrino Experiment Simulation.
The Deep Pyramid Convolution Neural Network(DPCNN) Method is implemented to generate results approximate to the dataset performance limit.
Two versions of Google Colab Notebooks are designed, optimized and simplified for further research and development, along with trained neural network model related. The notebooks are developed to run online, but they also support other notebook setups.
The default working directory is '/content/drive/My Drive'.
In the working directory,
Datasets should be found in './PhysicsData/'.
Models will be saved in './Neutrino_DPCNN_Model/'.
Answers will be produced in './DPCNN_Answers/'.