Technology-oriented research and innovation targeting the next generation of “Industrial Artificial Intelligence (AI)” can make a decisive contribution to maintaining and enhancing competitiveness. New generative AI technologies provide the opportunity to bring the data-driven models used for digitalization “to life” through machine learning and enable them to autonomously perform appropriate actions in order to relieve and support humans. The participation of companies, particularly SMEs, in dynamic and resilient value networks is essential for the European economy and for Germany as an export-oriented industrial location. This requires the rapid adaptation of all kinds of business processes, including logistics and production processes, to a wide range of existing and potentially new value-chain partners, while achieving the highest possible degree of automation in their execution. At the same time, applicable corporate policies and legal requirements must always be observed. These requirements, for example, help ensure fair competition and good working conditions while promoting the transition towards a sustainable circular economy.
In particular, the combination of the need for rapid adaptation to dynamic economic conditions, regulatory requirements, insufficient digitalization, and shortages of skilled personnel in these areas represents a major strategic challenge, both from the perspective of individual companies and from a broader economic perspective.
The LxM project is developing, on a prototype and pre-competitive basis, a multi-agent platform consisting of specialized agents for the (partially) autonomous planning and execution of business processes, with particular emphasis on compliance with legal and business rules and regulations, while appropriately involving humans in their role as decision-makers. To this end, Large Language Models are combined with neuro-symbolic AI methods and tools, and an integrated information architecture comprising Asset Administration Shells, knowledge graphs, and LLMs is created. The software developed will be evaluated in application use cases and, wherever possible, released as open-source software.
The project partners are:
- German Research Center for Artificial Intelligence
- Scheer IMC
- XPECT
Funding reference number: 16IS26030A
