When people talk about abstract topics, they often use linguistic imagery. Such metaphors make complex content easier to understand and provide clues as to how a topic is perceived. Prof. Dr. Ingrid Scharlau and Philip Porwol describe how metaphors shape our communication in an appealing way in their guidebook “Metaphors for Explaining”.
In their latest study, Tobias Peters and Professor Dr. Ingrid Scharlau of Project C01 examined how people deal with incorrect recommendations from AI systems and whether deliberately fostering distrust can improve users’ performance when using these systems. The key finding: calling for skepticism does not improve performance; instead, it tends to worsen it.
Can artificial intelligence work with doctors as equals and help them to make better diagnoses? A research team from the C05 project at TRR 318 is looking into this question. The computer scientists are working on an interactive system that assists doctors in making a diagnosis and checks and weighs up assumptions in dialog with them.
TRR 318 will be represented at the 2025 World Conference on Explainable Artificial Intelligence (XAI) in Istanbul from July 9 to 11. The conference brings together international experts every year to discuss the latest developments in XAI. This year, researchers from projects A03 and C02 will present their current work.
On 17 and 18 June, around 80 researchers with international and interdisciplinary perspectives came together for the TRR 318 conference “Contextualizing Explanations” in Bielefeld. The invited speakers and visitors discussed the importance of context in explanatory situations.
How do children learn their first words and how can care robots acquire the skills to help patients? In the fifth episode of the podcast “Explaining Explainability”, experts shed light on the concept of scaffolding - the art of supporting learners so that they can master complex tasks independently - and apply it to explanatory situations.
The current issue of the TRR 318 newsletter explores the role of context in AI explanations, aligning with the theme of the third TRR 318 conference “Contextualizing Explanations”, scheduled for June 17 and 18 at Bielefeld University.
Researchers from the INF project have developed a model that aims to predict the optimal prompts. This should improve the generated output of language models. Maximilian Spliethöver presented the TRR 318 study at the NAACL 2025 conference in New Mexico, USA.
Hubert Baniecki is a PhD student at the University of Warsaw and was a visiting researcher in project C03 of TRR 318 at the LMU Munich in March. In the interview, he shares what he researched with the TRR members, what made the collaboration special for him, and the impressions he took away.
Have we already found a solution with ChatGPT and DeepSeek for an explainable AI that meets our requirements for an AI that can explain itself? Moderator Prof. Dr.-Ing. Britta Wrede discusses this question with her guests Prof. Dr. Henning Wachsmuth and Prof. Dr. Axel Ngonga Ngomo in the fourth episode of the podcast “Explaining Explainability”.
How can explanations of AI systems be made comprehensible and what role does context play in this? These questions are the focus of the 3rd TRR 318 conference “Contextualizing Explanations”, which will take place on 17 and 18 June in Bielefeld.
From left: TRR 248 spokesperson Prof. Dr.-Ing Raimund Dachselt, Prof. Dr. Eyke Hüllermeier, Prof. Dr. Katharina Rohlfing, Prof. Dr. Kirsten Thommes, Prof. Dr. Philipp Cimiano, Prof. Dr.-Ing Anna-Lisa Vollmer and the project leaders of TRR 248, Prof. Dr. Vera Demberg and Prof. Dr. Markus Krötzsch.
On Thursday, a team of TRR 318 visited the Collaborative Research Center/Transregio 248 “Foundations of Perspicuous Software Systems”. The aim of the meeting was to identify interfaces between the two research networks and to intensify the scientific exchange.
The TRR 318 “Constructing Explainability” is now represented on Instagram. The new account sfb_trr318 offers regular updates and insights into interdisciplinary research on the topic of explainable artificial intelligence (XAI).
