We are looking for a postdoc who will work under the guidance of Prof. Vincent Aleven in the HCI Institute, on two sponsored projects, one with additional guidance provided by Dr. Steve Ritter at Carnegie Learning Inc., a Pittsburgh-based company. 

 

First, the postdoc will work on an IES-funded project, in which researchers from Carnegie Mellon University and Carnegie Learning, Inc. jointly address challenges teachers face in the context of mathematics learning in blended middle school classes that use intelligent tutoring software (ITS). As a next step in this project, we will run a classroom study in middle schools to rigorously evaluate whether and how adding decision support (a recommender system) to an analytics-based mixed-reality teacher support tool ("teacher smart glasses") affects teachers and students. The project will produce a new teacher orchestration tool, Lumilo 2, running on mixed-reality devices (smart glasses) so teachers can keep their eye on the class. The project will produce new scientific insight into how real-time analytics-based tools affect teacher help and, in turn, student learning. The project may lead to a more personalized K-12 math classroom.

 

In a second project, funded by the NSF, we are investigating how to address the challenge that intelligent tutoring software (ITS), although often effective in enhancing student learning, is not easy to build. Specifically, we focus on how Generative AI can be used to dramatically speed up the more laborious part of ITS authoring, namely, the creation of a domain model, either in the form of a behavior graph or a rule-based model of problem-solving knowledge. We explore these issues within the context of the CTAT authoring tool.  The project will contribute to a better understanding of how LLMs can be used, with a human in the loop, to generate appealing educational content that is effective for all students. It will contribute to a better understanding of whether and how LLMs are capable of modeling and representing procedural knowledge in a transparent, symbolic, and human-readable format for instructional purposes. More efficient, affordable authoring will likely contribute to the spread of ITSs – so that, eventually, many more students will be able to take advantage of this highly effective technology.