In May 2026, Yu Wang presented a paper on predicting states of understanding in explanatory interactions using cognitive load-related linguistic cues at the Fifteenth Language Recourses and Evaluation Conference (LREC 2026) in Palma, Spain. The results highlight that linguistic cues, such as surprisal, syntactic complexity, and interactive gaze vary with the listeners’ (explainee) level of understanding.
Wang, Y., Türk, O., Grimminger, A., & Buschmeier, H. (2026). Predicting states of understanding in explanatory interactions using cognitive load-related linguistic cues. Proceedings of the 15th Language Resources and Evaluation Conference, 11368–11378. Predicting States of Understanding in Explanatory Interactions Using Cognitive Load-Related Linguistic Cues - LREC 2026 .
Stefan Lazarov is going to present a paper as a poster on the dynamic use of self-adaptors by the explainees at the Cognitive Science Society conference in Rio de Janeiro in July 2026. In this study, the explainees’ self-adaptors were analyzed with regard to the topic sequence (novel vs. reintroduced) and the interlocutors who raises the topic during the interaction. The results indicate that explainees produce proportionally more self-adaptors as some explanation topics get reintroduced more frequently than other, suggesting that these topic repetitions could be related to potential moments of non-understanding.
Lazarov, S., Türk, O., Wagner, P., & Grimminger, A. (2026). Self-adaptors while listening about novel and repeated topics in explanations. Proceedings of the Annual Meeting of the Cognitive Science Society. [manuscript submitted for publication].
Another key achievement of project A02 is a study led by Olcay Türk on the multimodal detection of backchannels expressing incomplete understanding which is going to be presented at Interspeech in September 2026 in Sidney. Using acoustic, head movement, and discourse-related data, the researchers managed to show that incongruent backchannels (e.g., responding “yeah” while understanding is incomplete) are typically characterized by more neutral head movement configurations and lower acoustic dynamism in comparison to congruent backchannels which are surrounded by higher dynamics of motion and acoustic features.
Türk, O., Lazarov, S., Wang, Y., Grimminger, A., Buschmeier, H., & Wagner, P. (2026). When “yeah” means “not quite”: Multimodal detection of backchannels expressing incomplete understanding. Interspeech 2026. [manuscript submitted for publication].