InGram: Inducti… (2023-05-31T00:00:00.000000Z)

TL;DR

An INductive knowledge GRAph eMbedding method, InGram, that can generate embeddings of new relations as well as new entities at inference time and outperforms 14 different state-of-the-art methods on varied inductive learning scenarios.

Abstract

Inductive knowledge graph completion has been considered as the task of predicting missing triplets between new entities that are not observed during training. While most inductive knowledge graph completion methods assume that all entities can be new, they do not allow new relations to appear at inference time. This restriction prohibits the existing methods from appropriately handling real-world knowledge graphs where new entities accompany new relations. In this paper, we propose an INductive knowledge GRAph eMbedding method, InGram, that can generate embeddings of new relations as well as new entities at inference time. Given a knowledge graph, we define a relation graph as a weighted graph consisting of relations and the affinity weights between them. Based on the relation graph and the original knowledge graph, InGram learns how to aggregate neighboring embeddings to generate relation and entity embeddings using an attention mechanism. Experimental results show that InGram outperforms 14 different state-of-the-art methods on varied inductive learning scenarios.

Authors

Jaejun Lee

1 papers

Chanyoung Chung

1 papers

Joyce Jiyoung Whang

1 papers

References57 items

Representation Learning on Hyper-Relational and Numeric Knowledge Graphs with Transformers

Why So Gullible? Enhancing the Robustness of Retrieval-Augmented Models against Counterfactual Noise

Learning Representations of Bi-Level Knowledge Graphs for Reasoning beyond Link Prediction

Weisfeiler and Leman Go Relational

RAILD: Towards Leveraging Relation Features for Inductive Link Prediction In Knowledge Graphs

Abstract

References57 items

TL;DR

Abstract

Authors

References57 items

Representation Learning on Hyper-Relational and Numeric Knowledge Graphs with Transformers

Why So Gullible? Enhancing the Robustness of Retrieval-Augmented Models against Counterfactual Noise

Learning Representations of Bi-Level Knowledge Graphs for Reasoning beyond Link Prediction

Weisfeiler and Leman Go Relational

RAILD: Towards Leveraging Relation Features for Inductive Link Prediction In Knowledge Graphs

TL;DR

Abstract

Authors

References57 items

Representation Learning on Hyper-Relational and Numeric Knowledge Graphs with Transformers

Why So Gullible? Enhancing the Robustness of Retrieval-Augmented Models against Counterfactual Noise

Learning Representations of Bi-Level Knowledge Graphs for Reasoning beyond Link Prediction

Weisfeiler and Leman Go Relational

RAILD: Towards Leveraging Relation Features for Inductive Link Prediction In Knowledge Graphs

Relational Message Passing for Fully Inductive Knowledge Graph Completion

Semantic Grasping Via a Knowledge Graph of Robotic Manipulation: A Graph Representation Learning Approach

Inductive Knowledge Graph Reasoning for Multi-batch Emerging Entities

Incorporating Context Graph with Logical Reasoning for Inductive Relation Prediction

Subgraph Neighboring Relations Infomax for Inductive Link Prediction on Knowledge Graphs

Exploring Relational Semantics for Inductive Knowledge Graph Completion

Cycle Representation Learning for Inductive Relation Prediction

Knowledge Graph Reasoning with Relational Digraph

Improving Inductive Link Prediction Using Hyper-Relational Facts

NodePiece: Compositional and Parameter-Efficient Representations of Large Knowledge Graphs

Neural Bellman-Ford Networks: A General Graph Neural Network Framework for Link Prediction

Do Transformers Really Perform Bad for Graph Representation?

How Attentive are Graph Attention Networks?

Inductive Relation Prediction by BERT

Topology-Aware Correlations Between Relations for Inductive Link Prediction in Knowledge Graphs

Communicative Message Passing for Inductive Relation Reasoning

Inductive Entity Representations from Text via Link Prediction

The Surprising Power of Graph Neural Networks with Random Node Initialization

Inductively Representing Out-of-Knowledge-Graph Entities by Optimal Estimation Under Translational Assumptions

Relational Message Passing for Knowledge Graph Completion

5分で分かる!? 有名論文ナナメ読み：Jacob Devlin et al. : BERT : Pre-training of Deep Bidirectional Transformers for Language Understanding

Random Features Strengthen Graph Neural Networks

A Survey on Knowledge Graphs: Representation, Acquisition, and Applications

Inductive Relation Prediction by Subgraph Reasoning

Composition-based Multi-Relational Graph Convolutional Networks

DRUM: End-To-End Differentiable Rule Mining On Knowledge Graphs

Learning Attention-based Embeddings for Relation Prediction in Knowledge Graphs

Logic Attention Based Neighborhood Aggregation for Inductive Knowledge Graph Embedding

How Powerful are Graph Neural Networks?

RotatE: Knowledge Graph Embedding by Relational Rotation in Complex Space

Knowledge Graph Embedding: A Survey of Approaches and Applications

Graph Attention Networks

DeepPath: A Reinforcement Learning Method for Knowledge Graph Reasoning

Knowledge Transfer for Out-of-Knowledge-Base Entities : A Graph Neural Network Approach

Attention is All you Need

Inductive Representation Learning on Large Graphs

Analogical Inference for Multi-relational Embeddings

Modeling Relational Data with Graph Convolutional Networks

Differentiable Learning of Logical Rules for Knowledge Base Reasoning

Semi-Supervised Classification with Graph Convolutional Networks

Deep Residual Learning for Image Recognition

Observed versus latent features for knowledge base and text inference

Embedding Entities and Relations for Learning and Inference in Knowledge Bases

Understanding the difficulty of training deep feedforward neural networks

StATIK: Structure and Text for Inductive Knowledge Graph Completion

Inductive Relation Prediction Using Analogy Subgraph Embeddings

INDIGO: GNN-Based Inductive Knowledge Graph Completion Using Pair-Wise Encoding

BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding

Rectifier Nonlinearities Improve Neural Network Acoustic Models

Etr ← Triplets in the giant connected component of Etr

In G ′ , uniformly sample ninf entities and form Vinf by taking the sampled entities and their two-hop neighbors. We select at most 50 neighbors per entity for each hop to prevent exponential growth

Algorithm 2 shows how we generate the datasets used in Section 6. Also, Table 5 and Table 6 show the statistic of the datasets and the hyperparameters used to create the datasets

Field of Study

Journal Information

Name

Volume

Venue Information

Name

Type

URL

Alternate Names