Work Package 4
Data-Driven Use-Cases towards Exascale
Leader:
Dr. Maria Girone, CERN
This WP aims at the development and expansion of AI methods along representative use-cases from research and industry/SMEs, which have a strong focus on data-driven technologies, i.e., analyzing data-rich descriptions of physical phenomena. The structure of this WP is formulated such that the outcomes are applicable to intelligent workflows including innovative AI tool-chains, optimized on HPC-to-Exascale systems. The tasks contain the capabilities to evaluate prototype algorithms based on experimental and/or simulation data, code performance on Exascale HPC systems, and quality of data models.
The data models can support prediction and assessment of highly complex physical events and hidden risks. In the individual tasks, the management tools and the relevant methodologies concern data acquisition from multiple heterogeneous sources including traditional recorded acoustic/optical signals and videos, novel 3D visualization images, measurements by the HL-LHC, and large-scale simulation data. The knowledge discovered by data mining is able to detect patterns crossing the targeted database. These computational approaches employ HPC/HPDA/big data workflows, and Deep Learning analytic techniques to
-
generate essential information of the underlying dynamics in multidisciplinary systems
-
find new energy sources and improve the pipeline in storage and supply
-
identify unknown risks and the associated weakness in automation systems
-
obtain higher productivity, robust quality control, and cost-efficiency in manufacturing