In the CICERO project, our aim is to investigate if public speaking skills can be improved using virtual training. Descriptions of the user’s public speaking behavior are automatically extracted from audiovisual sensors, and an interactive virtual audience is used to provide feedback to the users depending on their public speaking performance. To investigate how virtual characters can help improving the learning outcome of a public speaking training system, I proposed a flexible and modular architecture for interactive virtual audiences, and conducted a series of studies on the impact of different feedback strategies on training outcomes and user experiences.
- A generic platform for training social skills with adaptative virtual agents
- Native vs. Non-native language fluency implications on multimodal interaction for interpersonal skills training
- Public speaking training with a multimodal interactive virtual audience framework
- An interactive virtual audience platform for public speaking training
- Influence of individual differences when training public speaking with virtual audiences