Confidence-aware Self-Semantic Distillation on Knowledge Graph Embedding

TL;DR

This paper introduces Confidence-aware Self-Knowledge Distillation (CSD), a novel method that improves low-dimensional Knowledge Graph Embedding by self-supervising from previous embeddings and filtering reliable knowledge without pre-trained teachers.

Contribution

The work proposes a simple, effective self-distillation approach with a semantic module to enhance low-dimensional KGE, avoiding complex operations and pre-training.

Findings

CSD improves performance across six KGE models and four datasets.

The semantic module effectively filters reliable knowledge.

CSD reduces computational costs compared to existing methods.

Abstract

Knowledge Graph Embedding (KGE), which projects entities and relations into continuous vector spaces, has garnered significant attention. Although high-dimensional KGE methods offer better performance, they come at the expense of significant computation and memory overheads. Decreasing embedding dimensions significantly deteriorates model performance. While several recent efforts utilize knowledge distillation or non-Euclidean representation learning to augment the effectiveness of low-dimensional KGE, they either necessitate a pre-trained high-dimensional teacher model or involve complex non-Euclidean operations, thereby incurring considerable additional computational costs. To address this, this work proposes Confidence-aware Self-Knowledge Distillation (CSD) that learns from the model itself to enhance KGE in a low-dimensional space. Specifically, CSD extracts knowledge from embeddings in previous iterations, which would be utilized to supervise the learning of the model in the next iterations. Moreover, a specific semantic module is developed to filter reliable knowledge by estimating the confidence of previously learned embeddings. This straightforward strategy bypasses the need for time-consuming pre-training of teacher models and can be integrated into various KGE methods to improve their performance. Our comprehensive experiments on six KGE backbones and four datasets underscore the effectiveness of the proposed CSD.