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.

Building Memory with Concept Learning Capabilities from Large-scale Knowledge Base

Continual knowledge infusion into pre-trained biomedical language models

Neural Axiom Network for Knowledge Graph Reasoning

Enhancing Key-Value Memory Neural Networks for Knowledge Based Question Answering.

Reasoning Through Memorization: Nearest Neighbor Knowledge Graph Embeddings

Neural-Symbolic Entangled Framework for Complex Query Answering

Concept learning through deep reinforcement learning with memory-augmented neural networks

Scalable Neural Methods for Reasoning With a Symbolic Knowledge Base

Knowledgebra: An Algebraic Learning Framework for Knowledge Graph

Towards flexible perception with visual memory

Question Answering over Knowledge Base using Language Model Embeddings

Revisiting Knowledge Base Embedding as Tensor Decomposition

KM 4 : Visual reasoning via Knowledge Embedding Memory Model with Mutual Modulation

Exploring knowledge graph-based neural-symbolic system from application perspective

Knowledge Enhanced Contextual Word Representations

Using Dynamic Semantic Network to Construct Knowledge Base

Knowledge Graph Embedding with Entity Neighbors and Deep Memory Network

Joint Knowledge Base Embedding with Neighborhood Context

Concept2Box: Joint Geometric Embeddings for Learning Two-View Knowledge Graphs

Neural Concept Binder

Knowledge-based end-to-end memory networks