Opening Data Silos In Administration For Usage In Data Science And Machine Learning

Practical example and current challenges

• Currently, mainly hyperscalers are being considered for the development of solutions for data science or machine learning use cases and the associated data storage. Digital sovereignty is therefore only possible to a limited extent. Public administration data is not available for all areas or not in suitable quality, or it is particularly protected and cannot be made available as raw da-ta. Instead, public administration data must be processed or aggregated so that start-ups or companies can use it.
• The Data Science & Machine Learning (DSML) Hub stands for openness and use of public admin-istration data (in consideration of data protection) in order to make it available for the develop-ment of data science and AI applications. The aim is to create a sustainable solution approach to the challenges of using data science and AI in the public sector. In addition, the hub will support companies, start-ups and scientific and research institutions in innovating and creating value through data science and AI.

1. The DSML Hub consists of five building blocks:
2. unlocking the administration's data silos and access to the data for third parties.
3. building an infrastructure for data science and AI with state-of-the-art hardware and common software (including a DSML platform)
4. expansion of the ecosystem currently being built up, consisting of start-ups, public IT service providers, (regional small and medium-sized) enterprises, administration, politics and scientific and research institutions
5. synchronisation with GAIA-X
6. implementation of ten selected innovative projects in administration and companies.

• The DSML platform with the corresponding infrastructure enables access to data and thus the processing of any use cases. Solutions can be developed for one's own needs or for third parties and transferred into operation. The offer is primarily addressed to regional companies, public administration and scientific and research institutions.
• The DSML Hub will strengthen Germany as a business and science location and create data spac-es that, in conjunction with the technical infrastructure, will make it possible to transfer data sci-ence and AI solutions from research or start-ups to regional companies.

What added value does the "GAIA-X project" offer?

• GAIA-X sets the stage for trust in data security and digital sovereignty.
• The Data Science & Machine Learning Hub provides the range of services required for develop-ment projects: configurable hardware (e.g. GPU, CPU, RAM) and a software stack (e.g. AI frameworks such as TensorFlow, Jupyter Notebook or tools for the data lifecycle). In addition, it will also be possible to offer pre-configured solutions within the framework of a marketplace. This makes it possible to carry out a development project up to the point of operation on the hub.
• GAIA-X enables national and international cooperation and increases the reach of the DSML Hub.

Use Case Team

• Tina Siegfried – Dataport
• Nikolai Wilckens – Dataport