Supporting Discursive Instruction Online and In the Classroom with Intelligent Conversational Agents

Speaker
Carolyn P. Rosé
Associate Professor of Language Technologies and Human-Computer Interactions, Carnegie Mellon University

When
-

Where
Newell-Simon Hall 1305 (Michael Mauldin Auditorium)

Video
Video link

Description

In the past six years, technology for dynamic support for collaborative learning has matured both in terms of its ability to monitor collaboration through automatic collaborative learning process analysis as well as to offer context appropriate support for effective participation in groups, such as using conversational agent technology. In early studies, these computer agents served the purpose of elevating the conceptual depth of collaborative discussions by leading students in groups through directed lines of reasoning, referred to as knowledge construction dialogues, that were meant to scaffold the process of groups constructing conceptually rich explanations together. Results from controlled studies demonstrate learning gains of up to 1.24 standard deviations in comparison with competitive control conditions. The academically productive talk form of classroom interaction is one in which a facilitator (or an agent) poses a question that calls for a relatively elaborated response (e.g., both a solution and a reason for the solution) and then presses the group to build on or challenge these ideas, with the purpose of keeping student reasoning at center stage and increasing student ownership of ideas. In the past two years we have pursued this research as part of a partnership with an urban school district, where we have run a teacher professional development program in which teachers were coached to use “academically productive talk” practices in their classrooms and online collaborative learning activities were used to prepare students for these whole class, teacher lead discussions using the same paradigm. The findings from a series of studies embedded within the two year professional development program provides evidence that novel conversational agent designs based on the “academically productive talk” approach to discussion facilitation improve learning during the online exercises, better prepare students to benefit from whole class discussions, and have a facilitating effect on the instructors who lead the whole class discussions.

Speaker's Bio

Dr. Rosé is an Associate Professor of Language Technologies and Human-Computer Interaction in the School of Computer Science at Carnegie Mellon University. Her research program is focused on better understanding the social and pragmatic nature of conversation, and using this understanding to build computational systems that can improve the efficacy of conversation between people, and between people and computers. In order to pursue these goals, she invokes approaches from computational discourse analysis and text mining, conversational agents, and computer supported collaborative learning. She serves on the executive committee of the Pittsburgh Science of Learning Center and the co-leader of its Social and Communicative Factors of Learning research thrust. She also serves as the Secretary/Treasurer of the International Society of the Learning Sciences and Associate Editor of the International Journal of Computer Supported Collaborative Learning.

Speaker's Website
http://www.cs.cmu.edu/~cprose/

Host
Brad Myers