Contributor
Vladimir Ilievski

Convolutional Deep Neural Networks on GPUs for Particle Physics Applications


Mentors
Sergei Gleyzer, Lorenzo Moneta, Enric Tejedor, Omar Zapata
Organization
CERN-HSF

The Convolutional Neural Networks (CNNs) are one special type of a deep learning neural networks with an enormous discriminative power for image classification. In fact, they significantly outperform the standard Computer Vision techniques of manually extracting image features and then building classifiers on top of them.

Recently, the physicist started to harness the power of the CNNs in the task of particle tagging. In their paper Oliveira et al. 2015 used the jet images in combination with CNNs to obtain better results to jet tagging in comparison to the standard physically-motivated features. Moreover, using the simulations of particle smashing, which in fact produces an accurate 3D image, is another potential use of the CNNs.

Since ROOT is the state-of-the-art data analysis tool extensively used in the High Energy Physics, it is of paramount importance to integrate a CNN implementation in its submodule called TMVA. For this reason, throughout this project I will implement such a solution.