Abstract:Knowledge graph is of great research value to artificial intelligence,which has been extensively applied in the fields of semantic search and question answering,etc.Knowledge graph representation transforms a large-scale knowledge graph comprising entities and relations into a continuous vector space.To this end,there have been a number of models and methods proposed for knowledge embedding.Among them,TransE is a classic translation-based method that is of low model complexity,high computational efficiency,as well as good capability of expressing knowledge.However,TransE still has two flaws:one is that it utilizes inflexible Euclidean distance as metric,and treats each feature dimension identically,hence,the model accuracy may be interfered by irrelevant dimensions;the other is that it has limitations in dealing with complex relations including reflexive,one-to-many,many-to-one and many-to-many relations.Currently,there has not been a single method that resolves the flaws simultaneously,and thus,we propose a revised translation-based method for knowledge graph representation,namely,TransAH.For the first flaw,TransAH adopts an adaptive metric,replacing Euclidean distance with weighted Euclidean distance by adding a diagonal weight matrix,which assigns different weights to every feature dimension.As to the second,inspired by TransH,it introduces the relation-oriented hyperspace model,projecting head and tail entities to hyperspace of a given relation for distinction.At last,empirical studies on public real knowledge graph datasets analyze and verify the effectiveness of the proposed method.Comprehensive comparative experiments using two tasks-link prediction and triplet classification show that,in contrast to the existing models and methods,TransAH achieves remarkable improvement in various aspects and demonstrates its superiority.

A Neural Translating General Hyperplane for Knowledge Graph Embedding.

Knowledge Graph Embedding by Translating on Hyperplanes

Knowledge Graph Embedding for Hyper-Relational Data

Incorporating Domain and Range of Relations for Knowledge Graph Completion.

Knowledge Graph Completion for Hyper-relational Data

Translation-Based Embeddings with Octonion for Knowledge Graph Completion

GTrans: Generic Knowledge Graph Embedding Via Multi-State Entities and Dynamic Relation Spaces

Transg : A Generative Model For Knowledge Graph Embedding

Learning Attention-Based Translational Knowledge Graph Embedding via Nonlinear Dynamic Mapping

Modeling Complex Relationship Paths for Knowledge Graph Completion

A Revised Translation-Based Method for Knowledge Graph Representation

TransA: an Adaptive Approach for Knowledge Graph Embedding

Text-Enhanced and Relational Context Based Hyperbolic Knowledge Graph Embedding

A Knowledge Graph Embedding Method Based on Neural Network

Generalizing Hyperedge Expansion for Hyper-relational Knowledge Graph Modeling

Knowledge Graph Embedding by Flexible Translation

Hyperbolic Hypergraph Neural Networks for Multi-Relational Knowledge Hypergraph Representation

TransE-MTP: A New Representation Learning Method for Knowledge Graph Embedding with Multi-Translation Principles and TransE

TransMS: Knowledge Graph Embedding for Complex Relations by Multidirectional Semantics

Translating Embeddings for Knowledge Graph Completion with Relation Attention Mechanism.

Efficient Parallel Translating Embedding For Knowledge Graphs