Experiment Description
The Munich Innovation Hub for Applied AI successfully addressed complex challenges in aeronautics, automotive, energy, manufacturing, and retail through the deployment of innovative AI solutions. The key outcome was the development and implementation of the Platform for Intuitive, Accountable AI (PIANAI), designed to enhance the efficiency, reliability, and intuitiveness of human-CPS interactions and cope with the rapid evolution of AI systems.
PIANAI, conceived as an overlay of the DIHIWARE and AI4EU platforms, was aligned with the AI Manufacturing Testbed provided by TNO. This alignment focused on International Data Spaces (IDS), GAIA-X, and federated learning, ensuring seamless integration of PIANAI services into the testbed. A notable application was the implementation of the “clearing house” concept, as defined by IDS.
Furthermore, the experiment offered a unique technical service that provided verifiable claims about the design, deployment, and use of AI services. These claims, supported by tamper-proof facts throughout the AI service's development process, ensured transparency and reliability. Additionally, the service allowed for post-design guarantees, offering insights into AI services from a black box perspective. This comprehensive approach not only enhanced the dependability of AI systems in these critical sectors but also set a precedent for future AI innovations and applications.
List Title
Impact
The experiment, comprising a series of smaller experiments, has made significant contributions to the DIH4AI platform through the integration of PIANAI as a technology service. Firstly, it has successfully integrated four regional AI assets into the DIH4AI platform, which is renowned for its on-demand innovative AI solutions "made in Europe." Secondly, the experiment has focused on designing functionalities in line with IDS and GAIA-X principles, leading to the effective integration of the regional platform PIANAI. This included addressing cloud-edge interoperability issues through containerized services.
Thirdly, the experiment has played a crucial role in fostering innovation and collaboration among DIHs by providing two key assets accessible via the DIH4AI platform: the PIANAI service and a federated learning service. Fourthly, it has served as a foundation for operationalizing AI applications' ethical assessment and certification, assessing four AI applications and providing verifiable claims for their auditing.
Furthermore, the Intra-DIH Regional experiment involved two SMEs, with the potential to engage up to four more in follow-up activities. Lastly, the experiment included cross-DIH collaborative activities, particularly in joint provision and development in partnership with TNO and their testbed. This comprehensive approach has not only bolstered the capabilities of the DIH4AI platform but also advanced the field of AI technology services, enhancing collaboration and ethical standards within the industry.