Abstract:Capturing the composition patterns of relations is a vital task in knowledge graph completion. It also serves as a fundamental step towards multi-hop reasoning over learned knowledge. Previously, several rotation-based translational methods have been developed to model composite relations using the product of a series of complex-valued diagonal matrices. However, these methods tend to make several oversimplified assumptions on the composite relations, e.g., forcing them to be commutative, independent from entities and lacking semantic hierarchy. To systematically tackle these problems, we have developed a novel knowledge graph embedding method, named DensE, to provide an improved modeling scheme for the complex composition patterns of relations. In particular, our method decomposes each relation into an SO(3) group-based rotation operator and a scaling operator in the three dimensional (3-D) Euclidean space. This design principle leads to several advantages of our method: (1) For composite relations, the corresponding diagonal relation matrices can be non-commutative, reflecting a predominant scenario in real world applications; (2) Our model preserves the natural interaction between relational operations and entity embeddings; (3) The scaling operation provides the modeling power for the intrinsic semantic hierarchical structure of entities; (4) The enhanced expressiveness of DensE is achieved with high computational efficiency in terms of both parameter size and training time; and (5) Modeling entities in Euclidean space instead of quaternion space keeps the direct geometrical interpretations of relational patterns. Experimental results on multiple benchmark knowledge graphs show that DensE outperforms the current state-of-the-art models for missing link prediction, especially on composite relations.

DEER: Descriptive Knowledge Graph for Explaining Entity Relationships

Open Relation Extraction Based on Core Dependency Phrase Clustering

ENT-DESC: Entity Description Generation by Exploring Knowledge Graph

Open Relation Modeling: Learning to Define Relations between Entities

Distributed Representation Learning for Knowledge Graphs with Entity Descriptions

Knowledge Graph Embedding with Diversity of Structures

Representation Learning of Knowledge Graphs with Entity Descriptions

Knowledge graph representation learning method and system in combination with entity description

Knowledge Representation Learning with Entities, Attributes and Relations.

Knowledge Graph Embeddings: A Comprehensive Survey on Capturing Relation Properties

A Comprehensive Study on Knowledge Graph Embedding over Relational Patterns Based on Rule Learning

DensE: An Enhanced Non-commutative Representation for Knowledge Graph Embedding with Adaptive Semantic Hierarchy

Entity-Relation Extraction As Full Shallow Semantic Dependency Parsing

Software Knowledge Entity Relation Extraction with Entity-Aware and Syntactic Dependency Structure Information

A Novel Table-to-Graph Generation Approach for Document-Level Joint Entity and Relation Extraction

A Cascade Dual-Decoder Model for Joint Entity and Relation Extraction

A Representation Learning Method of Knowledge Graph Integrating Ordered Relation Path and Entity Description Information

Augmenting Embedding Projection With Entity Descriptions for Knowledge Graph Completion

A Survey Deep Learning Based Relation Extraction

GraphER: A Structure-aware Text-to-Graph Model for Entity and Relation Extraction

VER: Unifying Verbalizing Entities and Relations