Concept Prerequisite Relation Prediction by Using Permutation-Equivariant Directed Graph Neural Networks

Xiran Qu, Xuequn Shang, Yupei Zhang

Track: Innovations in AI for Education (Day 1)

Paper Length: long-paper (6 pages + references)

Keywords: Concept Prerequisite Relation, permutation-equivariant GNNs, Weisfeiler- Lehman Test, Directed Graph Learning, AI for Education.

TL;DR: This paper proposes a directed graph neural network based on the Weisfeiler-Leman algorithm to address the concept prerequisite relation prediction, which is a fundamental task in using AI for education.

Abstract: This paper studies the problem of CPRP, concept prerequisite relation prediction, which is a fundamental task in using AI for education. CPRP is usually formulated into a link- prediction task on a relationship graph of concepts and solved by training the graph neural network (GNN) model. However, current directed GNNs fail to manage graph isomor- phism which refers to the invariance of non-isomorphic graphs, reducing the expressivity of resulting representations. We present a permutation-equivariant directed GNN model by introducing the Weisfeiler-Lehman test into directed GNN learning. Our method is then used for CPRP and evaluated on three public datasets. The experimental results show that our model delivers better prediction performance than the state-of-the-art methods.

Submission Number: 30