Link to CIFAR-10 example (regression) --> https://github.com/jglombitza/cifar_tutorial

Air shower reconstruction using deep neural networks

Train neural networks to reconstruct air-shower properties using detector responses measured at a hypothetic cosmic-ray observatory located at a high of 1400 m. The observatory features a cartesian array of 14 x 14 particle detectors with a distance of 750 m.

Each particle detector measures two quantities that are stored in the form of a cartesian image (2D array with 14 x 14 pixels). We will use these images to train neural networks to reconstruct the energy of the events.

Tutorial

We will use jupyter notebooks in the tutorial. As deep learning framework Keras is used.

For training the DNNs, we use Google colab to accelerate the training using a GPU. For opening the jupyter in colab, just click on the respective badge.

You can access the slides for the tutorial at:

Find more examples for deep learning in physics at www.deeplearningphysics.org/.

Neural Network Playground

Open the neural network playground at: https://playground.tensorflow.org and train your first neural network

Fully-connected network

Train a fully-connected network to reconstruct the energy of a cosmic-ray-induced air shower.

Convolutional neural network

Train a convolutional neural network to reconstruct the energy of a cosmic-ray-induced air shower.