Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Generalizing Hyperedge Expansion for Hyper-relational Knowledge Graph Modeling

Yu Liu, Shu Yang, Jingtao Ding, Quanming Yao, Yong Li

TL;DR

This paper generalizes the hyperedge expansion in hypergraph learning and proposes an equivalent transformation for HKG modeling, referred to as TransEQ, which considers both semantic and structural characteristics of a HKG to a KG, which considers both semantic and structural characteristics.

Abstract

By representing knowledge in a primary triple associated with additional attribute-value qualifiers, hyper-relational knowledge graph (HKG) that generalizes triple-based knowledge graph (KG) has been attracting research attention recently. Compared with KG, HKG is enriched with the semantic qualifiers as well as the hyper-relational graph structure. However, to model HKG, existing studies mainly focus on either semantic information or structural information therein, which however fail to capture both simultaneously. To tackle this issue, in this paper, we generalize the hyperedge expansion in hypergraph learning and propose an equivalent transformation for HKG modeling, referred to as TransEQ. Specifically, the equivalent transformation transforms a HKG to a KG, which considers both semantic and structural characteristics. Then an encoder-decoder framework is developed to bridge the modeling research between KG and HKG. In the encoder part, KG-based graph neural networks are leveraged for structural modeling; while in the decoder part, various HKG-based scoring functions are exploited for semantic modeling. Especially, we design the sharing embedding mechanism in the encoder-decoder framework with semantic relatedness captured. We further theoretically prove that TransEQ preserves complete information in the equivalent transformation, and also achieves full expressivity. Finally, extensive experiments on three benchmarks demonstrate the superior performance of TransEQ in terms of both effectiveness and efficiency. On the largest benchmark WikiPeople, TransEQ significantly improves the state-of-the-art models by 15\% on MRR.

Generalizing Hyperedge Expansion for Hyper-relational Knowledge Graph Modeling

TL;DR

This paper generalizes the hyperedge expansion in hypergraph learning and proposes an equivalent transformation for HKG modeling, referred to as TransEQ, which considers both semantic and structural characteristics of a HKG to a KG, which considers both semantic and structural characteristics.

Abstract

By representing knowledge in a primary triple associated with additional attribute-value qualifiers, hyper-relational knowledge graph (HKG) that generalizes triple-based knowledge graph (KG) has been attracting research attention recently. Compared with KG, HKG is enriched with the semantic qualifiers as well as the hyper-relational graph structure. However, to model HKG, existing studies mainly focus on either semantic information or structural information therein, which however fail to capture both simultaneously. To tackle this issue, in this paper, we generalize the hyperedge expansion in hypergraph learning and propose an equivalent transformation for HKG modeling, referred to as TransEQ. Specifically, the equivalent transformation transforms a HKG to a KG, which considers both semantic and structural characteristics. Then an encoder-decoder framework is developed to bridge the modeling research between KG and HKG. In the encoder part, KG-based graph neural networks are leveraged for structural modeling; while in the decoder part, various HKG-based scoring functions are exploited for semantic modeling. Especially, we design the sharing embedding mechanism in the encoder-decoder framework with semantic relatedness captured. We further theoretically prove that TransEQ preserves complete information in the equivalent transformation, and also achieves full expressivity. Finally, extensive experiments on three benchmarks demonstrate the superior performance of TransEQ in terms of both effectiveness and efficiency. On the largest benchmark WikiPeople, TransEQ significantly improves the state-of-the-art models by 15\% on MRR.

Paper Structure

This paper contains 33 sections, 2 theorems, 3 equations, 8 figures, 6 tables, 3 algorithms.

Table of Contents

  1. Introduction
  2. Related Work
  3. Knowledge Graph (KG) Modeling
  4. Hyper-relational Knowledge Graph (HKG) Modeling
  5. Preliminaries
  6. Hypergraph & Hyperedge Expansion
  7. Hyper-relational Knowledge Graph
  8. Method
  9. The TransEQ Model
  10. One Stone: Equivalent Transformation
  11. Two Birds: Encoder-Decoder Framework
  12. Model Training
  13. Other Variants of Generalized Transformations
  14. Theoretical Understanding
  15. Complexity Analysis
  16. ...and 18 more sections

Key Result

Theorem 1

In the conversion from a HKG to a KG, the generalized transformation in TransEQ is an equivalent transformation and preserves the complete information, while other variants of transformations like plain transformation and clique-based transformations can lead to partial information loss.

Figures (8)

  • Figure 1: An example of a HKG including primary triples and attribute-value qualifiers.
  • Figure 2: The illustration of star expansion on a hyperedge.
  • Figure 3: The architecture of our proposed HKG modeling model TransEQ. An original HKG is transformed to a KG via the equivalent transformation, and then a GNN-based encoder and a SF-based decoder are leveraged for modeling structural information and semantic information, respectively.
  • Figure 4: The illustration of the equivalent transformation.
  • Figure 5: The illustration of other variants of transformations.
  • ...and 3 more figures

Theorems & Definitions (4)

  • Definition 1
  • Definition 2
  • Theorem 1
  • Theorem 2