Information Technology and machine learning applications
Task 5.1 Worldwide distributed computing. Implementation of a common interface such as the DIRAC Cloud Computing
Element, for integration of virtual machines into the MC factory and testing the PaaS/SaaS services in INFN, CNRS/
IN2P3 clouds. INFN, CNRS will provide cloud resources and technologies for access to cloud infrastructures. INFN,
KCL, CNRS, DESY will provide use cases and test workflow over the cloud.
Task 5.2 Advanced network solutions. We will perform a set of joint stress tests sending data from the main sites in
Japan to the European Data Centers of Belle II and Hyper-K in Italy, France, UK, Germany and others. A set of common
tools will be used to monitor the network links and the performance will be compared with the estimated requirements
of the two experiments.
Task 5.3 Machine learning for big data analysis. Develop and compare different applications of ML algorithms for various
data analysis tasks. Share tools and methodologies thanks to a unified platform. Hold bi-annual workshops including
hands-on tutorial sessions.
Task 5.4 Machine learning for real time applications. Test of the Xilinx Versal FPGA hardware platform with benchmarks
and study of possible realtime ML applications of pattern recognition in the DAQ and trigger systems for Belle II. Application of the Versal system to anomalous cluster detection in the Belle II PXD, to disentangle possible exotic
particle signals from background.