Carnegie Mellon University

CMU Campus

April 01, 2019

Wednesday April 10, 2019 Spring Discourse Workshop

Open Participation Day April 10

Funded by the National Science Foundation Grant ACI-1443068
Carnegie Mellon University, Gates-Hillman Center 8102


Carolyn Penstein Rose, Carnegie Mellon University
Amanda Godley, University of Pittsburgh
Sherice Clarke, University of California at San Diego
Sadhana Puntambekar, University of Wisconsin at Madison

Invited Analysts:

Sidney D’Mello, University of Colorado at Boulder
Allison Hall, University of Illinois at Chicago
Peter Hastings, Depaul University
Cassie Quigley, University of Pittsburgh
Alina Reznitskaya, Montclair State University
James Wynn, Carnegie Mellon University
Noreen Webb, University of California at Los Angeles


This workshop is meant to foster collaboration and capacity in the application of human language technologies (HLT) to the study and improvement of K-12 education.  

Current research, theory, and policy surrounding K-12 instruction in the United States highlight the role of classroom talk in instructional quality and student learning. However, most studies of productive classroom talk are small scale and use researcher-specific analytical codes, in part because of the difficulty and expense of collecting and analyzing enough discourse data to draw valid inferences about the relationship between talk and learning across classroom settings. Recent advances in human language technologies offer a promising opportunity to address these challenges through the automatic and semi-automatic computer-based analysis of multiparty talk. Our proposed workshop is a first step towards solving these issues. Our aim is to begin building a shared database of classroom discourse data (video, audio and transcripts) using the Discourse DB platform to support storage, analysis, and sharing of discourse data for research.

The workshop will take place at Carnegie Mellon University (Gates-Hillman Center 8102) over two days (April 10th and 11th), however, the second day is by invitation only.  The first day is open to anyone who is interested and free of charge.


Program and Schedule

On the first day of the workshop,  invited participants will report on analyses of a shared dataset of  classroom discourse through Discourse DB. We will use these presentations to ground discussions about analysis, comparing and contrasting empirical and theoretical foundations of classroom talk, as well as collaboratively brainstorming useful features of discourse databases and platforms at the workshop.

8:30 am Welcome and Workshop Vision (Amanda Godley)
9:00 am Shared Dataset Overview (Sadhana Puntambekar)
9:30 am Analysis 1: James Wynn/Peter Hastings
10:00 am Coffee Break
10:30 am Analysis 2: Sidney D’Mello/Allison Hall
11:00 am Analysis 3: Cassie Quigley/Noreen Webb
11:30 am Analysis 4: Amanda Godley/Alina Reznitskaya
12:00 pm Discussant (Sherice Clarke)
12:30 pm Lunch and Networking
1:30 pm Hand Coding vs Auto Analysis (Sadhana Puntambekar and Carolyn Rose)
2:30 pm LightSIDE and DiscourseDB Tutorial (Carolyn Rose)
3:30 pm Coffee Break
4:00 pm Hands on work with DiscourseDB/LightSIDE
5:30 pm Open participation day closing