From the research laboratories to the factory floor or the power plant: AI technologies are increasingly being deployed in practice – whether in industrial production, the energy industry or the healthcare sector. This is accompanied by promises ranging from more efficient utilisation of resources to better products. At the same time, the actual use of AI in real environments harbours new challenges. One of the main hurdles lies in the handling of data: Inappropriate data for training AI models or poor data planning and descriptions lead to unexpected expenses or poor results. Finally, the development of AI models for a laboratory environment differs drastically from the development of models that have to perform in the real world. In addition, the traceability of AI continues to place high demands: Are the data sets in use actually suitable for the task at hand – and how can this be checked? Can users really understand the AI's predictions?

This is where the EU Horizon project AI-DAPT comes in, pursuing a data-centred approach. The objective is to strengthen the trustworthiness of AI solutions for their practical application. Under the leadership of the “Athena Research Centre”, 18 project partners research hybrid AI solutions that combine data-driven AI with scientific and theoretical expertise.

Solutions for intelligent data AI pipelines

The AI-DAPT project partners develop data-centred solutions for the entire lifecycle of artificial intelligence in productive environments. These enable an end-to-end automation across the entire cycle – from design to execution to evaluation of intelligent data AI pipelines that continuously learn and adapt to their context. The pipelines include steps such as data acquisition and data analysis as well as AI model training and deployment.

With the help of AI-DAPT solutions, continuous improvements to AI are also possible during ongoing operations: for example, the training of AI models can be adapted without interruption. The overarching objective is to ensure that artificial intelligence works reliably in different production environments. Specifically, AI-DAPT develops and links a series of interoperable services that are deployed in a specially developed AI-DAPT platform – and can also be integrated into other data analysis platforms.

Users play an active role in the AI system

AI-DAPT will combine the interoperable services with targeted “human-in-the-loop” approaches. These ensure that people play an active role in the AI system, for example by checking data or taking decisions suggested by the AI. The XAI (explainable AI) technologies will make the AI's decisions understandable for users.

The AI-DAPT solutions will be tested under real-life conditions with demonstrators in four different sectors: healthcare, robotics, energy and manufacturing. Together with project partners, Fraunhofer FOKUS supports and develops services for the design and maintenance of the data, assists in the architectural development of the pipeline and develops standards for communication between the services.