My research work for Hyperkamiokande experiment on e-/gamma classification using deep learning.
Neutrinos are some particles that are emmited by high energy events like supernovas or solar flares.
The purpose is to design a neural network that can efficiently classify the neutrino interactions recolected by the Cherenkov radiation in some observatories around the world.
NOTE: Some files are not updated. ViTs are not added to this repository yet.
https://docs.google.com/presentation/d/1OT-5TFxtsiA70FW9trtBICqJ7KzIa-w9i4ArOQQTFBY/edit?usp=sharing
There are many files regarding raw data transformation, visualization and some tests for classification using convolutional neural networks using Keras/Tensorflow.