(EPSRC-funded) This project develops common-sense and visually enhanced natural language generation for robots and other agents, enabling real-time human-agent communication in dynamic environments. It integrates multimodal data, external knowledge and adaptive reasoning to improve collaboration, support social robotics, and create applications in healthcare, public assistance and education.

(EPSRC-funded) This project develops emotionally aware, engaging natural language generation methods for mental health personal assistants, aiming to support young people through accessible, stigma-free advice. It focuses on generating fluent, varied, and empathetic responses from limited domain data to improve trust, engagement, and mental well-being support.

(EPSRC-funded) This project develops a collaborative storytelling framework that enables robots and humans to jointly interpret physical environments. By combining video analysis, knowledge acquisition, and language grounding, it aims to help robots better understand human activities and acquire missing environmental knowledge through interaction with people.

(RSE funded) This network addresses the spread of fake news by connecting scientific evidence with policy action. It focuses on misinformation challenges highlighted during the pandemic and in Scotland’s digital ethics strategies, aiming to develop targeted, evidence-based interventions to improve fact-checking, media literacy, and public resilience against harmful misinformation.

(RSE funded) This project will build a global dataset of emotional cues expressed through physical touch, pressure, and movement using sensor-enhanced robots. By capturing cultural variation in touch-based emotional expression, it aims to support the design of more inclusive, emotionally aware robots that communicate appropriately with users from diverse backgrounds.