DRIVER-SEAT
DRIVER-SEAT, or the Dialogue Reinforcement Infrastructure for Volitional Exploratory Research - Soliciting Effective Actions from Teachers, will provide teachers with a quick and effective way to respond to student online homework. Similar to Google’s Smart Reply, which uses machine learning to let users send predictive or “suggested” human-like messages when responding to an email, DRIVER-SEAT will give teachers three suggested automated messages to respond to students’ math homework. The teacher can choose the most appropriate selection from the three choices to send to their student.
What do DRIVER-SEAT comments look like?
Students will receive a message that includes their teacher’s comment with some context showing what the teacher is referring to with their comment.
Teachers will help create our library by piloting a prototype system and selecting feedback to send their students. Library development will enable us to apply deep learning in an effort to discover how to help teachers efficiently reply to their students.
Detecting how students are feeling
DRIVER-SEAT used the detectors that Dr. Heffernan created in collaboration with Dr. Ryan Baker to predict which students are confused, bored or engaged. These detectors have been used to predict end of year test scores, if students will go to college and which students major in a stem field.
DRIVER-SEAT provides teachers with recommended messages to send to students who appear confused. We anticipate these messages will help get students back on track. An typical message may sound like, "It seems you are having difficulties on your homework. Please come see me after school so we can work through this problem set together."
For the current DRIVER-SEAT study, the project is funded by the National Science Foundation (NSF) under Grant Number 1822830.