Affect Detection and Analysis
Facial expression trackers output activations for facial action units (AUs) that comprise different affects. We compile patterns of AUs seen in related work as well as use factor analysis to search for affect categories implicit in our corpus.
Data - Video, interaction log, affect annotations
Methods - Facial expression analysis, factor analysis, regression analysis
- Nye, B. D., Karumbaiah, S., Tokel, S. T., Core, M. G., Stratou, G., Auerbach, D., & Georgila, K. (2018) Engaging with the Scenario: Affect and Facial Patterns from a Scenario-Based Intelligent Tutoring System. Proceedings of the 19th International Conference on Artificial Intelligence in Education (AIED). [pdf]
- Nye, B., Karumbaiah, S., Tokel, S. T., Core, M. G., Stratou, G., Auerbach, D., & Georgila, K. (2017) Analyzing Learner Affect in a Scenario-Based Intelligent Tutoring System. Proceedings of the 18th International Conference on Artificial Intelligence in Education (AIED). [pdf]
In a randomized controlled study, we investigate the impact of different types of affective messages delivered by pedagogical agents in an online math tutor on student behavior and affect.
Data - Interaction log, survey
Methods - Randomized controlled trails, partial correlations, markov chains
[git]
- Karumbaiah, S., Lizarralde, R., Allessio, D., Woolf, B.P., Arroyo, I., Wixon, N. (2017) Addressing Student Behavior and Affect with Empathy and Growth Mindset. Proceedings of the 10th International Conference on Educational Data Mining (EDM). [pdf]