I am a doctoral candidate at SFB TRR318 “Constructing Explainability”, a collaborative research center at Paderborn and Bielefeld University. My work lies in the Interdisciplinary intersection of Cognitive Science and Psycholinguistics. This includes statistical modeling and computational cognitive science.
I investigate multimodal processes in both humans and artificial agents to understand the desiderata of successful explanations. The core idea across my investigation is to integrate XAI techniques with interpretable cognitive/neurosymbolic models to enable holistic explanation generation.
The idea goes beyond applying ‘cognitively agnostic’ classical XAI techniques (such as LIME/SHAP, contrastive explanations) in isolation, and instead utilizes XAI methods within theories and cognitive models of (language) processing.
Methodologically, I study human–human explanation using eyetracking and psycholinguistic experimentation.
For human–AI teaming, I work with foundation models (prompting, feature extraction, fine-tuning), Human-Robot Interaction, and statistical modeling (mostly Bayesian) to quantify explanation practices.
Updates
Recent talk at IEEE ICDL
An example scenario of multimodal interaction within HRI
