Combining HPC, AI and RDM: Challenges and Approaches

The availability of large amount of digital data, computational power combined with new approaches of dataanalysis and so called artificial intelligence created the potential of new scientific discoveries in almost all scientific fields.However, to realize the potential a number of challenges need to be addressed. The amount, quality and accessibility of datahas to be suitable. In short, the FAIR principles need to be implemented in a proper manner and the meta data has to besuitable and sufficient for the selected machine learning method. Many of the challenges can only be addressed if the ITinfrastructure offers suitable and comprehensive support in the fields of HPC, AI, and RDM. This talk will present theapproach taken by RWTH Aachen University and its partners to address the challenges.
The approach includes 1. Engagement in several projects of the German Initiative for Research Data Management (NFDI)with a focus on engineering. 2. Creation of a data integration platform Coscine implementing the FAIR principles and FAIRDigital Object interfaces. 3. Process Mining technologies to discover similarities between data sets and reveal otherwiseundetected workflows 4. Data stewards in large, long time collaborative projects (CRCs) 5. Integration of HPC and RDMplatforms to exploit data analytics capabilities of HPC.
Besides the scientific content the talk will cover the technical, financial, and human resources of the activities.