Jeremias Thun

Model Reconciliation through Explainability and Collaborative Recovery in Assistive Robotics

RACCOON: Grounding Embodied Question-Answering with State Summaries from Existing Robot Modules

Explainability is vital for establishing user trust, also in robotics. Recently, foundation models (e.g. vision-language models, VLMs) fostered a wave of embodied agents that answer arbitrary queries about their environment and their interactions with it. However, naively prompting VLMs to answer qu