Behavioural cloning of teachers for automatic homework selection

4Citations
Citations of this article
14Readers
Mendeley users who have this article in their library.
Get full text

Abstract

We describe a machine-learning system for supporting teachers through the selection of homework assignments. Our system uses behavioural cloning of teacher activity to generate personalised homework assignments for students. Classroom use is then supported through additional mechanisms to combine these predictions into group assignments. We train and evaluate our system against 50,065 homework assignments collected over two years by the Isaac Physics platform. We use baseline policies incorporating expert curriculum knowledge for evaluation and find that our technique improves on the strongest baseline policy by 18.5% in Year 1 and by 13.3% in Year 2.

Cite

CITATION STYLE

APA

Moore, R., Caines, A., Rice, A., & Buttery, P. (2019). Behavioural cloning of teachers for automatic homework selection. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 11625 LNAI, pp. 333–344). Springer Verlag. https://doi.org/10.1007/978-3-030-23204-7_28

Register to see more suggestions

Mendeley helps you to discover research relevant for your work.

Already have an account?

Save time finding and organizing research with Mendeley

Sign up for free