Optimal Embedding Guided Negative Sample Generation for Knowledge Graph Link Prediction

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Takamoto, Makoto, Oñoro-Rubio, Daniel, Rim, Wiem Ben, Maruyama, Takashi, Kotnis, Bhushan
Formato: Preprint
Publicado: 2025
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866910903478255616
author Takamoto, Makoto
Oñoro-Rubio, Daniel
Rim, Wiem Ben
Maruyama, Takashi
Kotnis, Bhushan
author_facet Takamoto, Makoto
Oñoro-Rubio, Daniel
Rim, Wiem Ben
Maruyama, Takashi
Kotnis, Bhushan
contents Knowledge graph embedding (KGE) models encode the structural information of knowledge graphs to predicting new links. Effective training of these models requires distinguishing between positive and negative samples with high precision. Although prior research has shown that improving the quality of negative samples can significantly enhance model accuracy, identifying high-quality negative samples remains a challenging problem. This paper theoretically investigates the condition under which negative samples lead to optimal KG embedding and identifies a sufficient condition for an effective negative sample distribution. Based on this theoretical foundation, we propose \textbf{E}mbedding \textbf{MU}tation (\textsc{EMU}), a novel framework that \emph{generates} negative samples satisfying this condition, in contrast to conventional methods that focus on \emph{identifying} challenging negative samples within the training data. Importantly, the simplicity of \textsc{EMU} ensures seamless integration with existing KGE models and negative sampling methods. To evaluate its efficacy, we conducted comprehensive experiments across multiple datasets. The results consistently demonstrate significant improvements in link prediction performance across various KGE models and negative sampling methods. Notably, \textsc{EMU} enables performance improvements comparable to those achieved by models with embedding dimension five times larger. An implementation of the method and experiments are available at https://github.com/nec-research/EMU-KG.
format Preprint
id arxiv_https___arxiv_org_abs_2504_03327
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Optimal Embedding Guided Negative Sample Generation for Knowledge Graph Link Prediction
Takamoto, Makoto
Oñoro-Rubio, Daniel
Rim, Wiem Ben
Maruyama, Takashi
Kotnis, Bhushan
Machine Learning
Computation and Language
Information Retrieval
Knowledge graph embedding (KGE) models encode the structural information of knowledge graphs to predicting new links. Effective training of these models requires distinguishing between positive and negative samples with high precision. Although prior research has shown that improving the quality of negative samples can significantly enhance model accuracy, identifying high-quality negative samples remains a challenging problem. This paper theoretically investigates the condition under which negative samples lead to optimal KG embedding and identifies a sufficient condition for an effective negative sample distribution. Based on this theoretical foundation, we propose \textbf{E}mbedding \textbf{MU}tation (\textsc{EMU}), a novel framework that \emph{generates} negative samples satisfying this condition, in contrast to conventional methods that focus on \emph{identifying} challenging negative samples within the training data. Importantly, the simplicity of \textsc{EMU} ensures seamless integration with existing KGE models and negative sampling methods. To evaluate its efficacy, we conducted comprehensive experiments across multiple datasets. The results consistently demonstrate significant improvements in link prediction performance across various KGE models and negative sampling methods. Notably, \textsc{EMU} enables performance improvements comparable to those achieved by models with embedding dimension five times larger. An implementation of the method and experiments are available at https://github.com/nec-research/EMU-KG.
title Optimal Embedding Guided Negative Sample Generation for Knowledge Graph Link Prediction
topic Machine Learning
Computation and Language
Information Retrieval
url https://arxiv.org/abs/2504.03327