How can artificial intelligence (AI) become comprehensible for industrial processes? This question is being investigated by the EU-funded project XMANAI (Explainable Manufacturing Artificial Intelligence). Fraunhofer FOKUS is participating as a scientific partner for real-time data management.

AI technologies are considered to be a major driver of innovation for society, science and economy. In industrial production, AI has the potential to make manufacturing faster, better and cheaper. However, nowadays self-learning systems are untransparent and not sufficiently comprehensible for stakeholders from industry and business. The EU-funded XMANAI project aims to contribute to a movement of AI systems from the “Black Box” to the “White Box”. The aim is to make machine learning explainable and to focus on both systems and data. The XMANAI data platform will provide important approaches to solving this problem.

Fraunhofer FOKUS is the scientific partner for real-time data management

In this project, Fraunhofer FOKUS is responsible for the work package “Asset Management Bundles Methods and System Designs”. Here, our scientists deal with real-time data acquisition and management, as well as with data protection and services for data storage at the same time. The main components we will design and develop are services for the:

- Collection and storage of data and data models,

- Preprocessing and acquisition of missing information,

- Securing of data provenance, as well as the securing of IPR for data production.

Contributing knowledge to data management ecosystem

FOKUS expert Dr. Yury Glikman, who is leading the work package, explains, “Data must be accurate, up-to-date and complete across the entire manufacturing process for AI systems to be successful and also to become traceable. To explore how this can best be achieved, we are bringing our knowledge around the data management ecosystem piveau® to the project”. The data platform developed by Fraunhofer FOKUS will represent the complete lifecycle of data and will be deployed to capture, aggregate and render it available for use. The platform will also be used to assess data quality, as well as to manage metadata.

The research project is funded by the European Commission under the Horizon 2020 research program. The project coordinator is the Italian company TXT E-SOLUTIONS S.P.A.