Carolyn Saund

62 Hillhead Street 

Room 574


University email

Personal Website


Research title: Modelling the Relationship Between Gesture Motion and Meaning

Research Summary

My research focuses on the generation of socially compelling non-verbal behavior on embodied agents. Specifically, I develop novel methods to produce communicatively meaningful complex gestures on virtual agents. This work draws from cognitive sciences, research and statistical methods in psychology, and machine learning and AI techniques including deep learning techniques, as well as inspiration and applications in digital and performance art. 

More broadly, I am interested in the social-communicative role of non-verbal behavior in specific situations, e.g. the role of facial expression in affect prediction at a sports match. 

Presently my projects focus on evaluating realtime gesture generation algorithms for semantic relevance and social impression. 


Saund, C., Roth, M., Chollet, M., & Marsella, S. (2019, September). Multiple metaphors in metaphoric gesturing. In 2019 8th International Conference on Affective Computing and Intelligent Interaction (ACII) (pp. 524-530). IEEE.


Saund, C., Birladeanu, A., & Marsella, S. (2021). CMCF: An architecture for realtime gesture generation by clustering gestures by motion and communicative function.


Saund, C., & Marsella, S. (2021, December). The Importance of Qualitative Elements in Subjective Evaluation of Semantic Gestures. In 2021 16th IEEE International Conference on Automatic Face and Gesture Recognition (FG 2021) (pp. 1-8). IEEE.


Saund, C., & Marsella, S. (2021). Gesture Generation. In The Handbook on Socially Interactive Agents: 20 years of Research on Embodied Conversational Agents, Intelligent Virtual Agents, and Social Robotics Volume 1: Methods, Behavior, Cognition (pp. 213-258).