Dynamic Support for Virtual Math Teams

Funded by: the National Science Foundation
PIs: Carolyn P. Rose and Gerry Stahl
Students, Staff, and Collaborators: Iris Howley, David Adamson, Hye-Ju Jang
Website: http://www.cs.cmu.edu/~cprose/VMT.html

Project Overview

This project builds on two prior SGER grants, which produced respectively a text mining tool bench called TagHelper tools (Rosť et al., 2008) (subsequently the LightSIDE toolbench (Mayfield & Rosť, in press)) and an architecture for managing heterogeneous teams online consisting of multiple humans and one or more intelligent conversational agents, originally called Basilica (Kumar & Rosť, 2011), now called Bazaar (Adamson & Rosť, 2012). These tools form the foundation for the emerging research area of dynamic support for online collaborative learning as well as assessment of learning processes in groups and dyads interacting face to face (Gweon et al., 2011; Gweon et al., 2012; Jain et al., 2012). PI Rosť has received two awards and several additional award nominations at a variety of HCI and Learning Sciences conferences for this work. Under this grant and synergistic funding, both tools have been substantially extended and used together in a series of over a dozen successful studies of high school and college level engineering, science and math learning (with pre to post test gains on assessments designed by course instructors) (e.g., Chaudhuri et al., 2009; Kumar et al., 2010; Ai et al., 2010; Kumar et al., 2011; Dyke et al., 2012; Howley et al., 2012), one study of group decision making (with an assessment of constructed plan quality using a rubric validated by the Navy used as a validation) (Kumar, 2011), and two completed PhD dissertations (Kumar, 2011; Gweon, 2012). Typical effect sizes reported in published work validating the approach are on the order of .5 to 1.25 standard deviations, which are considered medium to large effect sizes. In addition, TagHelper and LightSIDE are both used as learning tools in applied machine learning and introductory computational linguistics courses at Carnegie Mellon University and abroad. All tool benches produced with this funding are freely available for research purposes. Cumulatively, LightSIDE and TagHelper have a user base of several thousand users in 75 countries.


  1. Adamson, D. & Rosť, C. P. (2012). Coordinating Multi-Dimensional Support in Conversational Agents, in Proceedings of Intelligent Tutoring Systems
  2. Ai, H., Kumar, R., Nguyen, D., Nagasunder, A., Rosť, C. P. (2010). Exploring the Effectiveness of Social Capabilities and Goal Alignment in Computer Supported Collaborative Learning, in Proceedings of Intelligent Tutoring Systems.
  3. Chaudhuri, S., Kumar, R., Howley, I., Rosť, C. P. (2009). Engaging Collaborative Learners with Helping Agents, Proceedings of Artificial Intelligence in Education
  4. Dyke, G., Howley, I., Adamson, D., Rosť, C. P. (2012). Towards Academically Productive Talk Supported by Conversational Agents, in Proceedings of Intelligent Tutoring Systems
  5. Gweon, G., Agarwal, P., Udani, M., Raj., B., Rosť, C. P.(2011). The Automatic Assessmnet of Knowledge Integration Processes in Project Teams, in Proceedings of Computer Supported Collaborative Learning
  6. Gweon, G., Jain, M., McDonogh, J., Raj, B., Rosť, C. P. (2012). Predicting Idea Co-Construction in Speech Data using Insights from Sociolinguistics, in Proceedings of the International Conference of the Learning Sciences.
  7. Gweon, G. (2012). Assessment and support of the idea co-construction process that influences collaboration. Carnegie Mellon University, School of Computer Science, PhD dissertation.
  8. Howley, I., Adamson, D., Dyke, G., Mayfiled, E., Beuth, J., & Rosť, C. P. (2012). Group Composition and Intelligent Dialogue Tutors for Impacting Studentsí Self-Efficacy, in Proceedings of Intelligent Tutoring Systems
  9. Jain, M., McDonogh, J., Gweon, G., Raj, B., Rosť, C. P. (2012). An Unsupervised Dynamic Bayesian Network Approach to Measuring Speech Style Accommodation, in the Proceedings of the European Association for Computational Linguistics
  10. Kumar, R., Ai, H., Beuth, J., Rosť, C. P. (2010). Socially-capable Conversational Tutors can be Effective in Collaborative Learning Situations, in Proceedings of Intelligent Tutoring Systems.
  11. Kumar, R. (2011). Socially Capable Conversational Agents for Multi-Party Interactive Situations, Carnegie Mellon University, School of Computer Science, PhD dissertation.
  12. Kumar, R. & Rosť, C. P. (2011). Architecture for building Conversational Agents that support Collaborative Learning, IEEE Transactions on Learning 4(1), pp 21-34
  13. Kumar, R., Beuth, J., Rosť, C. P. (2011). Conversational Strategies that Support Idea Generation Productivity in Groups, in Proceedings of Computer Supported Collaborative Learning
  14. Mayfield, E. & Rosť, C. P. (in press). LightSIDE: Open Source Machine Learning for Text Accessible to Non-Experts, Invited chapter in the Handbook of Automated Essay Grading.
  15. Rosť, C. P., Wang, Y.C., Cui, Y., Arguello, J., Stegmann, K., Weinberger, A., Fischer, F., (2008). Analyzing Collaborative Learning Processes Automatically: Exploiting the Advances of Computational Linguistics in Computer-Supported Collaborative Learning, submitted to the International Journal of Computer Supported Collaborative Learning 3(3), pp237-271.