EduBERT: Pretrained Deep Language Models for Learning Analytics

TL;DR

This paper introduces EduBERT, a pretrained language model tailored for learning analytics, demonstrating improved performance on text classification tasks in educational data, with a smaller distilled version offering efficiency benefits.

Contribution

The paper presents EduBERT, a domain-specific pretrained language model for learning analytics, and shows its effectiveness over existing models, including a computationally efficient distilled version.

Findings

Pretrained EduBERT outperforms state-of-the-art on three learning analytics tasks.

A distilled version of EduBERT achieves comparable results with less computational cost.

Both models are publicly available for research use.

Abstract

The use of large pretrained neural networks to create contextualized word embeddings has drastically improved performance on several natural language processing (NLP) tasks. These computationally expensive models have begun to be applied to domain-specific NLP tasks such as re-hospitalization prediction from clinical notes. This paper demonstrates that using large pretrained models produces excellent results on common learning analytics tasks. Pre-training deep language models using student forum data from a wide array of online courses improves performance beyond the state of the art on three text classification tasks. We also show that a smaller, distilled version of our model produces the best results on two of the three tasks while limiting computational cost. We make both models available to the research community at large.