Publications and Projects

By themes / By date

Theme 1: Evaluate the Learning Impact of AI Errors in NLP-based Educational Systems in STEM

Summary: How do AI errors generated by NLP-based educational systems impact learning, and why? Do some learners get affected more? I perform empirical evaluations and propose a framework to assess such impacts. Projects on this topic have received funding support from Microsoft Research and Google.

Publications: (* = equal contribution)

Theme 2: Identify Learners’ Attitudinal and Perceptual Challenges with Imperfect and Opaque AI Educational Systems

Summary: What are learners’ perceptions, attitudes, and mental models towards an imperfect and opaque AI educational system, and why? So far, I have answered this question regarding a short-answer autograding system and a free-hand sketching feedback system. I identified various human-AI interaction challenges and proposed deployment guidelines. Projects on this topic have received funding support from Google.

Publications: (* = equal contribution)

Theme 3: Develop Novel AI Systems for Learning Contexts That Are Traditionally Hard to Scale

Context 1: Scaling open-ended non-STEM courses

Summary: Unlike many STEM courses, non-STEM courses tend to encourage introspection, critique, and subjective interpretation. Currently, however, large and online non-STEM courses are often taught similarly to STEM courses. How may AI help change the status quo? I performed user-centered ideations with stakeholders and proposed six future research directions.

Publications:

Context 2: Scaling spatial visualization training

Summary: How do we provide personalized and enjoyable spatial visualization training at scale? I developed a personalized feedback system for free-hand sketching problems and an online spatial visualization training platform deployed in a college classroom.

Publications:

Theme 4: Other HCI Topics

I have contributed my knowledge in Bayesian modeling in projects spanning a variety of HCI topics, including collective decision making, conversational agents, and social network analysis.

Publications: (* = equal contribution)