Representational Alignment Supports Effective Teaching

Ilia Sucholutsky, Katherine M. Collins, Maya Malaviya, Nori Jacoby, Weiyang Liu, Theodore R. Sumers, Michalis Korakakis, Umang Bhatt, Mark Ho, Joshua B. Tenenbaum, Zachary A. Pardos, Adrian Weller, Thomas L. Griffiths
Proceedings of the Innovation and Responsibility in AI-Supported Education Workshop, PMLR 273:146-173, 2025.

Abstract

A good teacher should not only be knowledgeable, but should also be able to communicate in a way that the student understands – to share the student’s representation of the world. In this work, we introduce a new controlled experimental setting, GRADE, to study pedagogy and representational alignment. We use GRADE through a series of machine-machine and machine-human teaching experiments to characterize a utility curve defining a relationship between representational alignment, teacher expertise, and student learning outcomes. We find that improved representational alignment with a student improves student learning outcomes (i.e., task accuracy), but that this effect is moderated by the size and representational diversity of the class being taught. We use these insights to design a preliminary classroom matching procedure, GRADE-Match, that optimizes the assignment of students to teachers. When designing machine teachers, our results suggest that it is important to focus not only on accuracy, but also on representational alignment with human learners.

Cite this Paper

BibTeX
@InProceedings{pmlr-v273-sucholutsky25a, title = {Representational Alignment Supports Effective Teaching}, author = {Sucholutsky, Ilia and Collins, Katherine M. and Malaviya, Maya and Jacoby, Nori and Liu, Weiyang and Sumers, Theodore R. and Korakakis, Michalis and Bhatt, Umang and Ho, Mark and Tenenbaum, Joshua B. and Pardos, Zachary A. and Weller, Adrian and Griffiths, Thomas L.}, booktitle = {Proceedings of the Innovation and Responsibility in AI-Supported Education Workshop}, pages = {146--173}, year = {2025}, editor = {Wang, Zichao and Woodhead, Simon and Ananda, Muktha and Mallick, Debshila Basu and Sharpnack, James and Burstein, Jill}, volume = {273}, series = {Proceedings of Machine Learning Research}, month = {03 Mar}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v273/main/assets/sucholutsky25a/sucholutsky25a.pdf}, url = {https://proceedings.mlr.press/v273/sucholutsky25a.html}, abstract = {A good teacher should not only be knowledgeable, but should also be able to communicate in a way that the student understands – to share the student’s representation of the world. In this work, we introduce a new controlled experimental setting, GRADE, to study pedagogy and representational alignment. We use GRADE through a series of machine-machine and machine-human teaching experiments to characterize a utility curve defining a relationship between representational alignment, teacher expertise, and student learning outcomes. We find that improved representational alignment with a student improves student learning outcomes (i.e., task accuracy), but that this effect is moderated by the size and representational diversity of the class being taught. We use these insights to design a preliminary classroom matching procedure, GRADE-Match, that optimizes the assignment of students to teachers. When designing machine teachers, our results suggest that it is important to focus not only on accuracy, but also on representational alignment with human learners.} }
Endnote
%0 Conference Paper %T Representational Alignment Supports Effective Teaching %A Ilia Sucholutsky %A Katherine M. Collins %A Maya Malaviya %A Nori Jacoby %A Weiyang Liu %A Theodore R. Sumers %A Michalis Korakakis %A Umang Bhatt %A Mark Ho %A Joshua B. Tenenbaum %A Zachary A. Pardos %A Adrian Weller %A Thomas L. Griffiths %B Proceedings of the Innovation and Responsibility in AI-Supported Education Workshop %C Proceedings of Machine Learning Research %D 2025 %E Zichao Wang %E Simon Woodhead %E Muktha Ananda %E Debshila Basu Mallick %E James Sharpnack %E Jill Burstein %F pmlr-v273-sucholutsky25a %I PMLR %P 146--173 %U https://proceedings.mlr.press/v273/sucholutsky25a.html %V 273 %X A good teacher should not only be knowledgeable, but should also be able to communicate in a way that the student understands – to share the student’s representation of the world. In this work, we introduce a new controlled experimental setting, GRADE, to study pedagogy and representational alignment. We use GRADE through a series of machine-machine and machine-human teaching experiments to characterize a utility curve defining a relationship between representational alignment, teacher expertise, and student learning outcomes. We find that improved representational alignment with a student improves student learning outcomes (i.e., task accuracy), but that this effect is moderated by the size and representational diversity of the class being taught. We use these insights to design a preliminary classroom matching procedure, GRADE-Match, that optimizes the assignment of students to teachers. When designing machine teachers, our results suggest that it is important to focus not only on accuracy, but also on representational alignment with human learners.
APA
Sucholutsky, I., Collins, K.M., Malaviya, M., Jacoby, N., Liu, W., Sumers, T.R., Korakakis, M., Bhatt, U., Ho, M., Tenenbaum, J.B., Pardos, Z.A., Weller, A. & Griffiths, T.L.. (2025). Representational Alignment Supports Effective Teaching. Proceedings of the Innovation and Responsibility in AI-Supported Education Workshop, in Proceedings of Machine Learning Research 273:146-173 Available from https://proceedings.mlr.press/v273/sucholutsky25a.html.

Related Material