by Garron Hillaire and Justin Reich
Teacher Moments (teachermoments.mit.edu) is an open-source platform for authoring digital clinical simulations. In Teacher Moments, participants are immersed in classroom vignettes through text, images, and video, and then called on to improvisationally respond through recorded audio and text. Artificially intelligent coaches can be programmed to review participant responses and provide feedback and scaffolding.
In this demo, we will first outline the research approach of working with 22 fellows comprised of teacher educators across the country and report on results of fostering collaboration for authoring, implementing, and researching the use of digital clinical simulations to prepare K-12 CS teachers to be equitable. Then we will highlight a specific simulation that leverages the extensible AI architecture in Teacher Moments to provide a dynamic support-based machine learning classifier that detects confusion in audio expression. We will report the predicted accuracy of the classifier and the preliminary results on how confusion relates to simulated difficult conversations on issues of equity in the classroom based on the intersection of cognitive theory and theory on learning from simulations. From the example simulation we will discuss two key concepts: 1) When should technology directly provide support to students learning about equitable teaching practice? 2) When should technology help facilitate teacher responses to students learning about equitable teaching practice?
