Ladies, if he:
He’s not your man, he’s a Deep Neural Network.
Creds: Alex Champandard/deeplearning.ai
Deep.TrkX (Deep Learning for Tracking) is a project launched at Uppsala University which is mandated to explore use of deep learning for event and track reconstruciton in PANDA Experiment (See FAIR). Some of the techniques being used are Dense Network (DNN), Covulutional (CNN)* and Recurrent Networks (RNN).
This is an HEP/ASCR DOE pilot project to evaluate and broaden the range of computational techniques and algorithms utilized in addressing HEP tracking challenges. Specifically the project will provide a framework to develop and evaluate new algorithms for track finding and classification, that will be demonstrated by applying advanced pattern recognition techniques to track candidate formation. For example, an optimized track formation algorithm that scales linearly with LHC luminosity, rather than quadratically or worse, may lead by itself to an order of magnitude improvement in the track processing throughput without affecting the track identification performance, hence maintaining the physics performance intact in the LHC upgrades. See more on official website HEP.TrkX.
