Incremental Knowledge Tracing from Multiple Schools

TL;DR

This paper investigates privacy-preserving knowledge tracing across multiple schools using continual learning, demonstrating that sequential learning with SAKT achieves performance comparable to pooled data models.

Contribution

It introduces a continual learning approach for knowledge tracing that respects data privacy policies across schools, using the ASSISTment dataset.

Findings

Sequential learning with SAKT performs similarly to pooled data models.

Privacy constraints can be addressed with continual learning in knowledge tracing.

The approach maintains high prediction accuracy without data sharing.

Abstract

Knowledge tracing is the task of predicting a learner's future performance based on the history of the learner's performance. Current knowledge tracing models are built based on an extensive set of data that are collected from multiple schools. However, it is impossible to pool learner's data from all schools, due to data privacy and PDPA policies. Hence, this paper explores the feasibility of building knowledge tracing models while preserving the privacy of learners' data within their respective schools. This study is conducted using part of the ASSISTment 2009 dataset, with data from multiple schools being treated as separate tasks in a continual learning framework. The results show that learning sequentially with the Self Attentive Knowledge Tracing (SAKT) algorithm is able to achieve considerably similar performance to that of pooling all the data together.