S · O Task Area 4: Transfer and Application

S · O dct:description

· O Transfer and Application aims to create a strong connection between the Data Science and AI sub-communities and NFDI4DS. We will focus on the sub-communities of (1) natural language processing and language technology as well as Semantic Web, (2) biomedical research and clinical decision-making, (3) information sciences and (4) social sciences. Additionally, we will activate other data-driven sub-communities by an open call for projects. The open call will allow for up to 30 speedboat projects whose feedback will influence our development road map and whose tangible results will be made available through our emerging infrastructure.

S · O Shared Task SOTA

S · O dct:description

· O Overview of shared task 2 of NFDI4DS

S · O Task Area 3: Infrastructure and Services

S · O dct:description

· O Infrastructure and Services establishes the infrastructure to collect and share all input, which we call Digital Objects (DO) which is required to deliver quality-assured data analytics solutions. The registries and repositories keep track of releases of quality assessed data as well as Data Science solutions required for assessment and the resulting benchmark information. The digital objects will be represented with rich metadata information and assembled as Research Knowledge Graphs for public use. Additional components ensure access to search and recommendation services (e.g., through a portal) and to additional services for public use such as data creation, annotation, and curation as well as an online authoring, publication, and execution platform. It further provides unified access to public HPC infrastructures

S · O Shared Task FORC

S · O dct:description

· O Overview of shared task 1 of NFDI4DS

S · O Task Area 1: Community and Training

S · O dct:description

· O Community and Training continuously analyzes requirements from the variety of stakeholder communities, including Computer Science research, Data Science and AI practitioners as well as industry. We develop and implement a comprehensive approach for DS and AI training, skill development, and capacity building, as well as leveraging existing platforms and resources. We comprehensively organize community building and stakeholder engagement using a variety of physical, online, and social networking channels with a special focus on facilitating diversity for example concerning gender balance and attracting young talents. We will furthermore increase awareness for legal and ethical aspects.

S · O Task Area 5: Interoperability and Cooperation

S · O dct:description

· O DS and AI involves a plethora of artifacts, e.g., datasets, models, ontologies, task definitions, code repositories, execution platforms, repositories, training materials, and so on. These artifacts are currently hidden in a number of platforms that manage the respective content. By making all digital artifacts available and interlinking them, NFDI4DS will foster interoperability, and collaboration between Data Science and AI platforms. We will collaborate with other NFDI consortia, as well as relevant national, European, and international partners, and with industry.

S · O Shared Task SOMD

S · O dct:description

· O Overview of shared task 3 of NFDI4DS

S · O Task Area 6: Management

S · O dct:description

· O Management aims to ensure a transparent and efficient management of the consortium while holding up to the highest legal and ethical standards. Efficient management and proper governance will be ensured within NFDI4DS. We will take care to comply with legal and contractual obligations and allocate funds properly. Meetings and decision processes will be organized and documented, and information will be distributed accordingly. NFDI4DS will publish relevant information and coordinate marketing activities.

S · O Task Area 2: Research Knowledge Graphs

S · O dct:description

· O Research Knowledge Graphs aims at improving FAIRness of Data Science artifacts including research datasets, benchmarks, machine learning models and research software (code and executables). The last years have seen a paradigm shift in Data and Computer Science towards data-driven and deep learning-based methods, which often rely on a combination of code, models, and underlying datasets. However, lack of transparency about data, code, or models is a cause for significant reproducibility and reusability issues, which surfaced across various domains. Therefore, we will follow an integrated approach towards representing and linking Data Science artifacts into a joint Research Knowledge Graph (RKG), enabling a transparent understanding of code/data provenance, model configurations, and training/testing data.