Fraunhofer FIZ for Hybrid Artificial Intelligence

Artificial Intelligence (AI) keeps changing the workspace, as well as the everyday life of countless people, across the globe more and more. Especially "Hybrid AI" is currently one of the most discussed topics in AI research. Symbolic AI, which works with high-level expert knowledge, meets data driven AI, using machine learning methods to independently deduce correlations from data sets. Combining these two fundamentally different models is the next biggest challenge in the development of artificial intelligence.

At the Research and Innovation Center for Hybrid AI, we are researching and developing technologies together with the Fraunhofer institute for Intelligent Analysis and Information Systems IAIS, that combine machine learning, global knowledge and logical reasoning. Our approaches to combine symbolic and data driven AI promise more efficient and easier usability, as well as a more understandable and transparent look into their functionality and results. Especially in the area of intelligent agents, which execute given tasks autonomously, the use of hybrid AI is essential.

• Hybrid AI

The unification of symbolic and data driven AI, which can interpret and understand meanings in a given context.

• Intelligent Agents

Autonomously working agents, which can communicate with humans, possess their own memory, use both models of generative AI as well as sensory tools, actuators and further abilities like programming, calculating or logical deduction.

• Generative AI

Development and usage of big language models, as well as adaptation to domain-specific contexts.

• Self-Learning Systems

Research and development of self-learning systems for demanding fields of application, supported by open-source models and technologies.