Robust Knowledge Graph Embedding via Denoising
This addresses robustness issues in knowledge graph embeddings for applications like recommendation systems or question answering, though it appears incremental as it builds on existing KGE methods.
The paper tackles the problem of making knowledge graph embeddings robust to perturbations in the embedding space, and the result is a framework that consistently outperforms state-of-the-art methods on benchmark datasets when faced with perturbed entity embeddings.
We focus on obtaining robust knowledge graph embedding under perturbation in the embedding space. To address these challenges, we introduce a novel framework, Robust Knowledge Graph Embedding via Denoising, which enhances the robustness of KGE models on noisy triples. By treating KGE methods as energy-based models, we leverage the established connection between denoising and score matching, enabling the training of a robust denoising KGE model. Furthermore, we propose certified robustness evaluation metrics for KGE methods based on the concept of randomized smoothing. Through comprehensive experiments on benchmark datasets, our framework consistently shows superior performance compared to existing state-of-the-art KGE methods when faced with perturbed entity embedding.