AI for Education
Building adaptive tutoring systems that respond to learners' cognitive load and support equitable access to personalized instruction.
Why this matters
Multimodal tutoring systems can improve learning outcomes, but poorly designed ones can also increase extraneous cognitive load — and the students most likely to benefit from personalized instruction are often the least well served by one-size-fits-all modality choices.
What we work on
- Real-time cognitive load estimation using response-time and interaction signals, without invasive instrumentation.
- Adaptive modality selection — choosing when to emphasize text, audio, or visual explanation based on a learner’s current state.
- Classroom-scale deployment and evaluation, in partnership with local school districts.
Representative outcomes
Across three classroom studies totaling 240 students, adaptive modality selection reduced extraneous cognitive load without harming learning gains — published in the International Journal of Artificial Intelligence in Education, and now the foundation of the EquiTutor project.
Back to all research
Abhiraj Bibhar
Department of Sociology and Anthropology
SRM University - Andhra Pradesh