Description

TensorFlow is a popular open-source software library for deep learning applications. It supports advanced math and flexible data-flows to build and train powerful deep neural networks.

Toolkit for Multivariate Analysis (TMVA) is a multi-purpose machine learning toolkit integrated into the ROOT scientific software framework, used in many particle physics data analyses and applications.

The goal of this project is to write an interface between TMVA and TensorFlow that allows use of TensorFlow’s functionality within TMVA.

Task ideas and expected results

• Fully-functional interface between Tensorflow and TMVA.
• It is important to design an interface that considers various input-data formats for relevant input data types: images, flat files, etc. on both sides.

Requirements: Strong C++ skills, good knowledge of deep learning, familiarity with Tensorflow, familiarity with ROOT and TMVA a plus.

Mentors:

• Sergei Gleyzer
• Fernanda Psihas
• Stefan Wunsch

Links:

• http://root.cern/tmva
• TMVA source code

Corresponding Project

• TMVA

Participating Organizations

• Florida
• CERN
• KIT
• UTA

---