Fraunhofer Institute of Optronics, System Technologies and Image Exploitation IOSB Our AI-based assistance systems support people by enabling them to process their tasks more effectively and efficiently. The systems can guide users through tasks, relieve them of routine activities, and even take on more complex tasks that they would never be able to complete themselves in this quantity and quality.
This applies in particular to the evaluation and analysis of large amounts of data and the resulting information, such as extensive quantities of images, videos, texts, or data series. AI can help to identify hidden patterns and correlations, draw further conclusions (reasoning), and integrate heterogeneous data and information into an overall picture, merge it, prepare it as required, and visualize it.
Decisions are usually based on a large amount of information that decision-makers are rarely able to perceive and interpret simultaneously. We develop interactive software systems for decision support that enable well-founded and traceable decisions to be made, taking all relevant aspects into account.
The learning ability of AI-based assistance systems enables them to continuously improve and adapt to changing requirements and new challenges. Their adaptability allows the systems to adapt to the individual needs and preferences of users and provide them with tailored support.
We use both data-driven (sub-symbolic) AI methods, which learn relevant correlations from large amounts of data, and knowledge-based (symbolic) AI methods, which map existing knowledge and thus enable a semantic understanding of data and information. We also combine these approaches in hybrid (e.g., neuro-symbolic) AI, which combines the strengths of both approaches by, for example, improving the interpretability of AI results.