Representational Alignment Supports Effective Machine Teaching
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2024
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866910812756508672 |
|---|---|
| author | Sucholutsky, Ilia Collins, Katherine M. Malaviya, Maya Jacoby, Nori Liu, Weiyang Sumers, Theodore R. Korakakis, Michalis Bhatt, Umang Ho, Mark Tenenbaum, Joshua B. Love, Brad Pardos, Zachary A. Weller, Adrian Griffiths, Thomas L. |
| author_facet | Sucholutsky, Ilia Collins, Katherine M. Malaviya, Maya Jacoby, Nori Liu, Weiyang Sumers, Theodore R. Korakakis, Michalis Bhatt, Umang Ho, Mark Tenenbaum, Joshua B. Love, Brad Pardos, Zachary A. Weller, Adrian Griffiths, Thomas L. |
| contents | 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. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2406_04302 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Representational Alignment Supports Effective Machine Teaching Sucholutsky, Ilia Collins, Katherine M. Malaviya, Maya Jacoby, Nori Liu, Weiyang Sumers, Theodore R. Korakakis, Michalis Bhatt, Umang Ho, Mark Tenenbaum, Joshua B. Love, Brad Pardos, Zachary A. Weller, Adrian Griffiths, Thomas L. Machine Learning 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. |
| title | Representational Alignment Supports Effective Machine Teaching |
| topic | Machine Learning |
| url | https://arxiv.org/abs/2406.04302 |