Abstract:Knowledge graph reasoning (KGR) aims to infer new knowledge or detect noises, which is essential for improving the quality of knowledge graphs. Recently, various KGR techniques, such as symbolic- and embedding-based methods, have been proposed and shown strong reasoning ability. Symbolic-based reasoning methods infer missing triples according to predefined rules or ontologies. Although rules and axioms have proven effective, it is difficult to obtain them. Embedding-based reasoning methods represent entities and relations as vectors, and complete KGs via vector computation. However, they mainly rely on structural information and ignore implicit axiom information not predefined in KGs but can be reflected in data. That is, each correct triple is also a logically consistent triple and satisfies all axioms. In this paper, we propose a novel NeuRal Axiom Network (NeuRAN) framework that combines explicit structural and implicit axiom information without introducing additional ontologies. Specifically, the framework consists of a KG embedding module that preserves the semantics of triples and five axiom modules that encode five kinds of implicit axioms. These axioms correspond to five typical object property expression axioms defined in OWL2, including ObjectPropertyDomain, ObjectPropertyRange, DisjointObjectProperties, IrreflexiveObjectProperty and AsymmetricObjectProperty. The KG embedding module and axiom modules compute the scores that the triple conforms to the semantics and the corresponding axioms, respectively. Compared with KG embedding models and CKRL, our method achieves comparable performance on noise detection and triple classification and achieves significant performance on link prediction. Compared with TransE and TransH, our method improves the link prediction performance on the Hits@1 metric by 22.0% and 20.8% on WN18RR-10% dataset, respectively.

Adapting Distilled Knowledge for Few-shot Relation Reasoning over Knowledge Graphs

Adapting Meta Knowledge Graph Information for Multi-Hop Reasoning over Few-Shot Relations

Multi-view Few-shot Reasoning for Emerging Entities in Knowledge Graphs

Few-shot Knowledge Graph Relational Reasoning via Subgraph Adaptation

Meta Relational Learning for Few-Shot Link Prediction in Knowledge Graphs

Relation Adversarial Network for Low Resource Knowledge Graph Completion

Cognitive Knowledge Graph Reasoning for One-shot Relational Learning.

Context-Aware Adapter Tuning for Few-Shot Relation Learning in Knowledge Graphs

Multi-Hop Knowledge Graph Reasoning in Few-Shot Scenarios

Attention-based Exploitation and Exploration Strategy for Multi-Hop Knowledge Graph Reasoning

Relational Learning with Gated and Attentive Neighbor Aggregator for Few-Shot Knowledge Graph Completion

Neural Axiom Network for Knowledge Graph Reasoning

Meta-Learning Based Dynamic Adaptive Relation Learning for Few-Shot Knowledge Graph Completion

Introducing Graph Neural Networks for Few-Shot Relation Prediction in Knowledge Graph Completion Task

Learning to Sample and Aggregate: Few-shot Reasoning over Temporal Knowledge Graphs

Relation-oriented few-shot knowledge graph prototype networks

Adaptive Attentional Network for Few-Shot Relational Learning of Knowledge Graphs

Learning Relation-Specific Representations for Few-shot Knowledge Graph Completion

Few-Shot Relation Extraction with Dual Graph Neural Network Interaction.

Stepwise relation prediction with dynamic reasoning network for multi-hop knowledge graph question answering

Intra-Graph-Based Few-Shot Link Prediction in Knowledge Graph