← All Research Groups
Information Profiling and Retrieval
The search for knowledge: efficient discovery of research findings with AI tools
The research group “Information Profiling and Retrieval”, headed by Professor Ralf Krestel, focuses on the optimum findability and accessibility of relevant information for research, such as scholarly publications. Using methods of Artifical Intelligence, information is processed, analysed, and presented in forms easily understood by users.
For this, the machine learning experts use methods of intelligent information processing and provision, including text mining, information retrieval, recommender systems, and natural language processing. With innovative methods and techniques, the research group aims to improve the information landscape and to make relevant content more accessible.
DID YOU KNOW?
The largest language model currently has 1.8 trillion parameters. By comparison: the human brain has 100 billion neurons with 100 trillion synapses.
Language models and knowledge graphs are the focus of ZBW research
The ZBW research group “Information Profiling and Retrieval” is also based on a professorship at the Faculty of Engineering at Kiel University and is currently focused on two key technologies. Language models, such as ChatGPT, offer an enormous store of knowledge. Such generative models are able to produce texts and thus to answer questions. Studying and evolving domain-specific models is of particular interest for the research group.
However, these statistical models are still prone to hallucinations, i.e. the model invents facts that are not true. A deeper understanding of such language models is therefore required for future successful application in the context of Open Science. Knowledge graphs, the second focus of the research group, are in some respect the counterpart to language models. They are not based on statistical analysis of gigantic text and data corpora, but are specifically fed with facts which they connect in a large graph. This enables the efficient storage and retrieval of knowledge. Combining these two technologies and their application offers great potential within the context of the ZBW.
Foto: Timo Wilke
“The fascination of my research field Information Profiling and Retrieval is the complexity which is needed to provide users with the information that is relevant to them. For this, documents and texts must be indexed semantically and processed individually for the users. Knowledge graphs and language models help store the world’s knowledge and react to users’ widely different information needs. With our work we can support people to find precisely the relevant information needed for their research, decision-making, and personal development. It’s exciting to continually evolve our techniques and methods in this dynamic field, and to adapt and use the latest AI methods, and thus to meet changing needs.”
Professor Ralf Krestel
Head of the research group “Information Profiling and Retrieval”
Open Science AT ZBW
Research groups and their current topical focus
Digital Information Infrastructures
Transparency and openness: new infrastructures for digital resources in research
Information Profiling and Retrieval
Language models and knowledge graphs are the focus of ZBW research
Web Science
Studying the impact and relevance of scientific publications
Digital Economics
Science about science: studying research texts and processes
Engagement
Science policy consulting at the ZBW
The developments associated with Open Science take place at many levels of the science system. As a Leibniz institution, it is a matter of concern for the ZBW to see its own experiences and research results inform the debate in science policy bodies at the national, European and international level.
vision
Future
Lorem ipsum dolor sit amet, consetetur sadipscing elitr, sed diam nonumy eirmod tempor invidunt ut labore et dolore magna aliquyam erat, sed diam voluptua. At vero eos et accusam et justo duo dolores et ea rebum. Stet clita kasd gubergren, no sea takimata sanctus est Lorem ipsum dolor sit amet. Lorem ipsum dolor sit amet, consetetur sadipscing elitr, sed diam nonumy eirmod tempor invidunt ut labore et dolore magna aliquyam erat, sed diam voluptua.
