| Title | Social Skills Training with Virtual Assistant and Real-Time Feedback |
| Publication Type | Conference Paper |
| Year of Publication | 2017 |
| Authors | Ali, Mohammad Rafayet, Hoque, Ehsan |
| Conference Name | Proceedings of the 2017 ACM International Joint Conference on Pervasive and Ubiquitous Computing and Proceedings of the 2017 ACM International Symposium on Wearable Computers |
| Publisher | ACM |
| Conference Location | New York, NY, USA |
| ISBN Number | 978-1-4503-5190-4 |
| Keywords | Human Behavior, human factors, nonverbal cues, pubcrawl, Scalability, Social Agents, social skills, virtual agent |
| Abstract | Nonverbal cues are considered the most important part in social communication. Many people desire people; but due to the stigma and unavailability of resources, they are unable to practice their social skills. In this work, we envision a virtual assistant that can give individuals real-time feedback on their smiles, eye-contact, body language and volume modulation that is available anytime, anywhere using a computer browser. To instantiate our idea, we have set up a Wizard-of-Oz study in the context of speed-dating with 47 individuals. We collected videos of the participants having a conversation with a virtual agent before and after of a speed-dating session. This study revealed that the participants who used our system improved their gesture in a face-to-face conversation. Our next goal is to explore different machine learning techniques on the facial and prosodic features to automatically generate feedback on the nonverbal cues. In addition, we want to explore different strategies of conveying real-time feedback that is non-threatening, repeatable, objective and more likely to transfer to a real-world conversation. |
| DOI | 10.1145/3123024.3123196 |
| Citation Key | ali_social_2017 |
Social Skills Training with Virtual Assistant and Real-Time Feedback