Modeling Relation Paths for Representation Learning of Knowledge Bases

Abstract

Representation learning of knowledge bases (KBs) aims to embed both entities and relations into a low-dimensional space. Most existing methods only consider direct relations in representation learning. We argue that multiple-step relation paths also contain rich inference patterns between entities, and propose a path-based representation learning model. This model considers relation paths as translations between entities for representation learning, and addresses two key challenges: (1) Since not all relation paths are reliable, we design a path-constraint resource allocation algorithm to measure the reliability of relation paths. (2) We represent relation paths via semantic composition of relation embeddings. Experimental results on real-world datasets show that, as compared with baselines, our model achieves significant and consistent improvements on knowledge base completion and relation extraction from text.

10 pages

References in corpus (2)

• Max-Margin Nonparametric Latent Feature Models for Link Prediction
• Learning New Facts From Knowledge Bases With Neural Tensor Networks and Semantic Word Vectors

Cited by in corpus (44)

• Unifying Knowledge Graph Learning and Recommendation: Towards a Better Understanding of User Preferences
• Analogical Inference for Multi-Relational Embeddings
• Differentiable Learning of Logical Rules for Knowledge Base Reasoning
• TransA: An Adaptive Approach for Knowledge Graph Embedding
• Modeling Relational Data with Graph Convolutional Networks
• On the representation and embedding of knowledge bases beyond binary relations
• TransG : A Generative Mixture Model for Knowledge Graph Embedding
• TransGCN:Coupling Transformation Assumptions with Graph Convolutional Networks for Link Prediction
• Neural Programmer: Inducing Latent Programs with Gradient Descent
• From One Point to A Manifold: Knowledge Graph Embedding For Precise Link Prediction
• RDF2Rules: Learning Rules from RDF Knowledge Bases by Mining Frequent Predicate Cycles
• Does William Shakespeare REALLY Write Hamlet? Knowledge Representation Learning with Confidence
• Knowledge Graph Embedding with Iterative Guidance from Soft Rules
• Knowledge Representation Learning: A Quantitative Review
• Faithful Embeddings for Knowledge Base Queries
• Iteratively Learning Embeddings and Rules for Knowledge Graph Reasoning
• One-Shot Relational Learning for Knowledge Graphs
• Knowledge Graph Embedding with Entity Neighbors and Deep Memory Network
• Incorporating Relation Paths in Neural Relation Extraction
• Interstellar: Searching Recurrent Architecture for Knowledge Graph Embedding
• Open-World Knowledge Graph Completion
• TorusE: Knowledge Graph Embedding on a Lie Group
• Learn to Explain Efficiently via Neural Logic Inductive Learning
• Knowledge-driven Site Selection via Urban Knowledge Graph
• Recurrent Skipping Networks for Entity Alignment
• MDE: Multiple Distance Embeddings for Link Prediction in Knowledge Graphs
• Predicting ConceptNet Path Quality Using Crowdsourced Assessments of Naturalness
• Rule-Guided Compositional Representation Learning on Knowledge Graphs
• Investigations on Knowledge Base Embedding for Relation Prediction and Extraction
• Logic-guided Semantic Representation Learning for Zero-Shot Relation Classification
• Recurrent One-Hop Predictions for Reasoning over Knowledge Graphs
• A Structural Representation Learning for Multi-relational Networks
• Important Attribute Identification in Knowledge Graph
• SCEF: A Support-Confidence-aware Embedding Framework for Knowledge Graph Refinement
• Distributional Negative Sampling for Knowledge Base Completion
• Semantically Enhanced Models for Commonsense Knowledge Acquisition
• Information Extraction from Scientific Literature for Method Recommendation
• Multi-task Neural Network for Non-discrete Attribute Prediction in Knowledge Graphs
• Expeditious Generation of Knowledge Graph Embeddings
• Improved Knowledge Base Completion by Path-Augmented TransR Model
• PPKE: Knowledge Representation Learning by Path-based Pre-training
• Convolutional Neural Knowledge Graph Learning
• Mining Commonsense Facts from the Physical World
• Semantic Vector Spaces for Broadening Consideration of Consequences