Information & Natural Language Processing

I-NLP stands for Information & Natural Language Processing. Our team researches and develops various computational linguistic methods, in particular AI-based information extraction (IE), and uses these in further stages of development to create knowledge-based, intelligent AI assistants. Our IE tools support expert-based research in extensive, heterogeneous data collections, while our assistance systems enable intuitive, natural access to complex engineering software (specialized software, engineering software) designed for experts.
For our research work, we use various language models (LLMs), embeddings, knowledge graphs, RAG and GraphRAG technologies (retrieval-augmented generation), as well as deep learning image and table transformers. 

Our research findings are incorporated into practical software tools 

• that can be used to analyze digital technical documents consisting of continuous text, lists, and complex tables, as well as structured SQL databases, and
• knowledge-based AI assistants that are based on reliable and accurate sources of information. They can be operated on-premises or in protected clouds – naturally in compliance with all relevant data protection guidelines.

An important quality feature of our assistance tools is source transparency, which links each generated answer to the underlying documents.
The focus has so far been on technical application domains such as energy management, IT management, building and process automation.

Projects that have been completed can be found on the German subpage Knowledge Grounded Dialog Systems.