Project A01: Adaptive explanation generation
In Project A01, researchers from linguistics, psychology, and computer science are studying two types of adaptation: cognitive adaptivity and interactive adaptivity. Cognitive adaptivity refers to how the explainer perceives their counterpart, and how this changes during the explanatory process. This is referred to as the partner model. Interactive adaptivity refers to how the explainer adjusts their form of explanations as the dialogue unfolds.
In the current funding phase, the team is complementing the partner model with a self-model, including one's current role in the dialogic explanation, and investigating the extent to which these are explicitly verbalized in interactive strategies (e.g., metacommunication). These processes of cognitive and interactive adaptivity are implemented in a model for computer-based generation of explanations in order to enable conversational XAI systems to respond flexibly and effectively to different explanation situations.
Research areas: Psychology, Computer science, Linguistics
Support Staff
Vincent Emmerling. Bielefeld University
Paula Herrmann, Paderborn University
Marie Weise, Paderborn University
Former Members
Amelie Robrecht, Research associate
Erick Kiprop Ronoh, Research associate
Posters
Conference-Poster presented at CogSci 2024 of the publication "Changes in Partner Models - Effects of Adaptivity in the Course of Explanations" by Heike M. Buhl, Josephine B. Fisher and Katharina J. Rohlfing.
Conference-Poster presented at 4th Summer School on Social Human-Robot Interaction with the title "Substantiveness in explanation" by Josephine B. Fisher.