TuckER: Tensor Factorization for Knowledge Graph Completion

Abstract

Knowledge graphs are structured representations of real world facts. However, they typically contain only a small subset of all possible facts. Link prediction is a task of inferring missing facts based on existing ones. We propose TuckER, a relatively straightforward but powerful linear model based on Tucker decomposition of the binary tensor representation of knowledge graph triples. TuckER outperforms previous state-of-the-art models across standard link prediction datasets, acting as a strong baseline for more elaborate models. We show that TuckER is a fully expressive model, derive sufficient bounds on its embedding dimensionalities and demonstrate that several previously introduced linear models can be viewed as special cases of TuckER.

Cited by in corpus (15)

• A Survey on Knowledge Graphs: Representation, Acquisition and Applications
• A Benchmarking Study of Embedding-based Entity Alignment for Knowledge Graphs
• Structure-Augmented Text Representation Learning for Efficient Knowledge Graph Completion
• Bringing Light Into the Dark: A Large-scale Evaluation of Knowledge Graph Embedding Models Under a Unified Framework
• IMF: Interactive Multimodal Fusion Model for Link Prediction
• A Review of Knowledge Graph Completion
• From Discrimination to Generation: Knowledge Graph Completion with Generative Transformer
• MulDE: Multi-teacher Knowledge Distillation for Low-dimensional Knowledge Graph Embeddings
• Link Prediction with Attention Applied on Multiple Knowledge Graph Embedding Models
• Rule Mining over Knowledge Graphs via Reinforcement Learning
• SAppKG: Mobile App Recommendation Using Knowledge Graph and Side Information-A Secure Framework
• Predicting affinity ties in a surname network
• Survey on Computational Applications of Tensor Network Simulations
• SPARK: Adaptive Low-Rank Knowledge Graph Modeling in Hybrid Geometric Spaces for Recommendation
• Textual Entailment for Effective Triple Validation in Object Prediction