AI data platform
The AI data platform of the Baden-Württemberg AI Alliance makes it easier for companies, start-ups and public institutions to provide and access AI models and corresponding quality-assured data sets, thereby promoting the development of innovative AI applications.
- Open source: design, implementation and provision of an operational, customizable platform
- Compliance by design: compatibility with ethical principles and legal framework conditions
- HPC resources: Access to the execution of AI models.
- Tools + Services: Automated data quality services check and optimize uploaded data sets for AI use.
Data for AI? Yes. But how?
Many companies and research institutions struggle with incorrect, incomplete or insufficiently annotated data. These shortcomings make it difficult to develop reliable AI solutions and slow down innovation. At the same time, the manual preparation of data is resource-intensive and expensive - especially for SMEs without specialized data science teams.
In practice, there is often a lack of:
- Search option for AI assets
- Systematic data management within and between companies
- Standards for data formats and quality
- sufficiently annotated training data for AI
- Specialists with expertise in AI and data management
In addition, many organizations are reluctant to share sensitive data because security and data sovereignty are not guaranteed.
Sector-specific data rooms
Sector-specific data spaces enable the standardized, secure and sovereign exchange of data within an industry - for example in mobility, healthcare or tourism. Through common rules, interfaces and metadata standards, they create trust, promote interoperability and facilitate the development and use of AI applications.
Production
For the data platform, we are developing solutions for the common good that give manufacturing companies secure access to production data.
By creating tools and technologies for connecting data sources, e.g. on the basis of the Industry 4.0 Asset Administration Shell, and supporting the use of existing data rooms (e.g. in accordance with the concepts and technologies of the Manufacturing-X program), we promote innovation and competitiveness in the industry.
Health
Health data rooms are currently emerging across Europe and are also leading to a rethink in the healthcare sector in Germany. The Freiburg University Medical Center is contributing its experience in the trustworthy, federated use of health data and the secure provision of data for the development of innovative medical products to the AI data platform. The aim of networking with the AI data platform is also to support interdisciplinary applications, for example in the field of urban resilience or environmental medicine, with data. Legal and ethical aspects of patient protection play a key role here.
Smart City
Together with Daten::Raum::Freiburg, we are working on opening up municipal data for AI use cases. The aim is to link data sets and create a blueprint for the broad use of municipal data in the AI ecosystem. The city of Freiburg is contributing its already well-developed data platform to the project and linking it with data from other sectors. The AI data platform is based on the piveau software, which is already an integral part of an urban data room architecture in Freiburg and is also used throughout Europe for public data portals. This will be further developed in the project for AI use in innovative ecosystems.
Mobility
Legally compliant synthetic data for autonomous driving, historical data sets for predictive models to support multi-modal mobility or data generation at the test bench: we are working on aggregating data and data services from a wide variety of sources to create a commercially viable offering in the field of mobility. Data-driven methods as a cornerstone for new, efficient mobility are integrated into the AI data platform by a network of partners from the Karlsruhe area.
We enable search options across data rooms
Application example: Intelligent visitor forecasts in tourism
In order to be able to predict when and how many people will be at a particular hotspot in tourism, a combination of different data sources is required - for example:
- Event dates (What takes place when?)
- Mobility data (How do people move through the region?)
- Weather forecast data (How does the weather affect visitor flows?)
A sector-specific data room for tourism creates the basis for bringing this data together securely and evaluating it with AI. This allows peak visitor numbers to be better predicted and managed in a targeted manner - for example through real-time recommendations, optimized traffic management or adapted infrastructure planning. This increases guest satisfaction, reduces the burden on the region and enables data-based decisions for local authorities and providers.
What is missing - and why is this a problem?
In AI development in particular, it is important to have quick access to high-quality data and tried-and-tested models - ideally domain-specific and compatible with in-house systems. However, without a central overview, many players lose time, money and innovation potential. Small and medium-sized enterprises (SMEs) in particular hardly have the resources to overcome these hurdles on their own.
Cross-sector offers
Many challenges in the field of artificial intelligence not only affect individual industries, but are also relevant across all sectors - such as the processing of large amounts of image data, compliance with data protection standards or the use of pre-trained models. This is precisely where cross-sector offerings come in.
Data quality
Data quality determines the success of AI. Incomplete, incorrect or poorly structured data leads to unreliable results. The AI data platform offers automated tools for checking, cleansing and annotating data - for better models and efficient processes. This turns raw data into quality-assured data. This lays the foundation for trustworthy AI applications in business, science and administration.
Ethics and law
Trustworthy AI needs ethical guard rails. Our platform integrates ethical principles right from the start - through transparent processes, traceable decisions and fair data processing. In this way, we ensure that AI applications are not only efficient, but also responsible and socially acceptable. Ethics is not an add-on, but a central building block of our digital future.
AI Usability
AI doesn't have to be complicated. Our platform places particular emphasis on user-friendliness: intuitive interfaces, clear instructions and context-based help make AI assets easily accessible - even without in-depth prior knowledge. In this way, we enable broad, practical access to AI technologies for companies, administration and research. Because good usability is the key to successful AI use.
AI computing cluster
Powerful computing resources for AI applications. The AI data platform provides access to high-performance computing (HPC) in order to efficiently train and operate even computationally intensive AI models. We deliberately rely on partners from Baden-Württembergwho provide scalable infrastructure, short computing times and secure data processing for research, industry and administration.
Advantages of the AI data platform
The AI data platform enables low-threshold access to AI assets (data and AI models) and thus establishes a market for AI assets. In this way, the development and application of innovative AI applications in various sectors in the business and science location of Baden-Württemberg is promoted and synergy effects between business, public institutions and research are maximized.
Become an early adopter in the AI data platform user forum!
Would you like to play an active role in shaping the development of a future-oriented AI ecosystem? Then the AI data platform user forum is the right place for you!
A metadata-based catalog system will be created by the end of 2025 as the first expansion stage. Operations are scheduled to start in 2026 in a public-private model - and you can be part of it from the start.Why take part?
As an early adopter in the user forum you will receive:
- Early access to developments and test phases
- Opportunities to help shape functions, standards and priorities
- Visibility on the platform with logo & statement
- Direct exchange with leading AI players from business, research and administration
Whether you want to provide data and models, develop an innovative application scenario or test the platform in practice: We look forward to your participation!
Your contact person
Do you have any questions or would you like to find out more about our AI solutions? Then feel free to send me an e-mail or give me a call - I look forward to hearing from you!
- Project manager of the AI data platform
- thomas.uslaender@iosb.fraunhofer.de
Our partners of the AI data platform:
Supported by
The KI-Allianz Baden-Württemberg eG and the associated projects are an initiative that aims to promote the development and application of artificial intelligence (AI). The measures are intended to help make Baden-Württemberg a hotspot for AI development and application, connect existing initiatives and drive forward the digital transformation in a supra-regional network.
