Tensor factorization for temporal knowledge graph forecasting

Manuel Dileo

doi:10.1016/j.neucom.2026.132846

Tensor factorization for temporal knowledge graph forecasting

2026-01-29

0

PRE

AI

DOI：10.1016/j.neucom.2026.132846

Original

Original request for help

Save

Abstract

En 中文

• We revisit tensor factorization for temporal knowledge graph forecasting. • We extend TNTComplEx with an RBF timestamp encoder and a temporal regularizer. • Our model achieves comparable or superior performance to state-of-the-art deep models. • We report 5 to 30 MRR improvements over previous tensor factorization baselines. • The method is substantially more efficient in training and inference time compared to state-of-the-art graph neural network models.

Tensor factorization for temporal knowledge graph forecasting

Abstract

Journal

Organization