AIED
When Features Misrepresent Underrepresented Learners: Auditing Algorithmic Bias with Differentially Expressive Features
Thu Jul 2, 2:45 PM–3:00 PM · North 206
Equity, Fairness & Inclusion Explainable & Trustworthy AI Learner & Student Modeling Educational Data Mining Methods
★ Notable speakers
Shamya Karumbaiah
★
— Affect detection, student engagement modeling, and equity/bias in intelligent tutoring and adaptive learning
Audits algorithmic bias in educational models by examining how features that misrepresent underrepresented learners introduce unfair disparities.