Classifying Math KCs via Task-Adaptive Pre-Trained BERT

TL;DR

This paper enhances automatic classification of educational content's knowledge components by expanding input types, increasing label granularity, and applying task-adaptive BERT, achieving higher accuracy and proposing a new evaluation measure.

Contribution

It introduces a novel approach using task-adaptive pre-trained BERT for multi-label classification of KCs with expanded inputs and labels, outperforming previous methods.

Findings

Prediction accuracy improved by 0.5-2.3% over baselines.

Expanded input types and labels increase practical applicability.

Proposed evaluation measure recovers 56-73% of mispredicted labels.

Abstract

Educational content labeled with proper knowledge components (KCs) are particularly useful to teachers or content organizers. However, manually labeling educational content is labor intensive and error-prone. To address this challenge, prior research proposed machine learning based solutions to auto-label educational content with limited success. In this work, we significantly improve prior research by (1) expanding the input types to include KC descriptions, instructional video titles, and problem descriptions (i.e., three types of prediction task), (2) doubling the granularity of the prediction from 198 to 385 KC labels (i.e., more practical setting but much harder multinomial classification problem), (3) improving the prediction accuracies by 0.5-2.3% using Task-adaptive Pre-trained BERT, outperforming six baselines, and (4) proposing a simple evaluation measure by which we can recover 56-73% of mispredicted KC labels. All codes and data sets in the experiments are available at:https://github.com/tbs17/TAPT-BERT