AI-SUPPORTED DISCOVERY OF MOTIVATIONAL MEASURES
In my dissertation project, I explore how to support a Human-AI collaborative process for building measurement models of student's motivations. Through analysis of simulated and real student data, I explore the information affordances of extending dichotomous item-response theory to the estimation of latent motivations from multi-measure behavioral scales that were automatically derived from data.
LIVE MEASUREMENT OF STUDENT MOTIVATIONS
In this project, I draw on theories of self-control and value-based decision making to tackle the challenges of measuring student motivations using natural observations of student behavior in log data. I have explored how to leverage information about the contexts in which particular student behaviors occur can be discriminative indicators of latent motivations.
DATA-DRIVEN DISCOVERY OF MODELS (D3M)
As part of the CMU D3M team, I designed and developed a tool that enables users with no data science experience to intuitively find answers to questions about their data. In this project, I experimented with how to expose advanced automatic machine learning technologies to untrained users through a block-based workflow tool.
PERSONALIZED LEARNING SQUARED
In this project, I used quantitative and qualitative research methods to support the development of an online application that supports educators in identifying and supporting the motivational and cognitve needs of students through analysis and monitoring of students usage of learning applications.
This project utilized user-centered research methods to explore unmet training needs of low-level managers from existing professional development offerings. As part of the project ,we developed ShareSight, a proof-of-concept online application that leveraged learner-sourcing mechanisms and reflection on everyday experiences to provide learners with more personally grounded practice of curriculum concepts.