Complementarity-driven Representation Learning for Multi-modal Knowledge Graph Completion

TL;DR

This paper introduces MoCME, a novel framework for multi-modal knowledge graph completion that leverages complementarity in multi-modal data and employs entropy-guided negative sampling, achieving state-of-the-art results.

Contribution

The paper proposes MoCME, a new framework that exploits intra- and inter-modal complementarity and uses entropy-guided negative sampling for improved multi-modal knowledge graph completion.

Findings

MoCME outperforms existing methods on five benchmark datasets.

The complementarity-guided fusion improves entity representations.

Entropy-guided negative sampling enhances training robustness.

Abstract

Multi-modal Knowledge Graph Completion (MMKGC) aims to uncover hidden world knowledge in multimodal knowledge graphs by leveraging both multimodal and structural entity information. However, the inherent imbalance in multimodal knowledge graphs, where modality distributions vary across entities, poses challenges in utilizing additional modality data for robust entity representation. Existing MMKGC methods typically rely on attention or gate-based fusion mechanisms but overlook complementarity contained in multi-modal data. In this paper, we propose a novel framework named Mixture of Complementary Modality Experts (MoCME), which consists of a Complementarity-guided Modality Knowledge Fusion (CMKF) module and an Entropy-guided Negative Sampling (EGNS) mechanism. The CMKF module exploits both intra-modal and inter-modal complementarity to fuse multi-view and multi-modal embeddings, enhancing representations of entities. Additionally, we introduce an Entropy-guided Negative Sampling mechanism to dynamically prioritize informative and uncertain negative samples to enhance training effectiveness and model robustness. Extensive experiments on five benchmark datasets demonstrate that our MoCME achieves state-of-the-art performance, surpassing existing approaches.