Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Revisit and Outstrip Entity Alignment: A Perspective of Generative Models

Lingbing Guo, Zhuo Chen, Jiaoyan Chen, Yin Fang, Wen Zhang, Huajun Chen

TL;DR

This work reframes embedding-based entity alignment (EEA) as a generative problem and introduces Generative EEA (GEEA) with a mutual variational autoencoder (M-VAE) to both align entities across knowledge graphs and synthesize new entities with concrete features. By decomposing the objective into reconstruction, distribution matching, and prediction components, the authors show that generative objectives can improve EEA performance while enabling unconditional and conditional entity synthesis. GEEA achieves state-of-the-art results on multiple multi-modal KG benchmarks and demonstrates high-quality generation of new entities, with a compact variant still delivering strong gains. The approach advances multi-modal KG embeddings and offers practical benefits for knowledge integration, virtual-world design, and downstream QA tasks.

Abstract

Recent embedding-based methods have achieved great successes in exploiting entity alignment from knowledge graph (KG) embeddings of multiple modalities. In this paper, we study embedding-based entity alignment (EEA) from a perspective of generative models. We show that EEA shares similarities with typical generative models and prove the effectiveness of the recently developed generative adversarial network (GAN)-based EEA methods theoretically. We then reveal that their incomplete objective limits the capacity on both entity alignment and entity synthesis (i.e., generating new entities). We mitigate this problem by introducing a generative EEA (GEEA) framework with the proposed mutual variational autoencoder (M-VAE) as the generative model. M-VAE enables entity conversion between KGs and generation of new entities from random noise vectors. We demonstrate the power of GEEA with theoretical analysis and empirical experiments on both entity alignment and entity synthesis tasks.

Revisit and Outstrip Entity Alignment: A Perspective of Generative Models

TL;DR

This work reframes embedding-based entity alignment (EEA) as a generative problem and introduces Generative EEA (GEEA) with a mutual variational autoencoder (M-VAE) to both align entities across knowledge graphs and synthesize new entities with concrete features. By decomposing the objective into reconstruction, distribution matching, and prediction components, the authors show that generative objectives can improve EEA performance while enabling unconditional and conditional entity synthesis. GEEA achieves state-of-the-art results on multiple multi-modal KG benchmarks and demonstrates high-quality generation of new entities, with a compact variant still delivering strong gains. The approach advances multi-modal KG embeddings and offers practical benefits for knowledge integration, virtual-world design, and downstream QA tasks.

Abstract

Recent embedding-based methods have achieved great successes in exploiting entity alignment from knowledge graph (KG) embeddings of multiple modalities. In this paper, we study embedding-based entity alignment (EEA) from a perspective of generative models. We show that EEA shares similarities with typical generative models and prove the effectiveness of the recently developed generative adversarial network (GAN)-based EEA methods theoretically. We then reveal that their incomplete objective limits the capacity on both entity alignment and entity synthesis (i.e., generating new entities). We mitigate this problem by introducing a generative EEA (GEEA) framework with the proposed mutual variational autoencoder (M-VAE) as the generative model. M-VAE enables entity conversion between KGs and generation of new entities from random noise vectors. We demonstrate the power of GEEA with theoretical analysis and empirical experiments on both entity alignment and entity synthesis tasks.
Paper Structure (36 sections, 2 theorems, 29 equations, 5 figures, 13 tables, 1 algorithm)

This paper contains 36 sections, 2 theorems, 29 equations, 5 figures, 13 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Revisit Embedding-based Entity Alignment
  3. Preliminaries
  4. Entity Alignment
  5. Entity Synthesis
  6. Generative Models
  7. EEA Benefits from the Generative Objectives
  8. The Limitations of GAN-based EEA Methods
  9. Generative Embedding-based Entity Alignment
  10. Mutual Variational Autoencoder
  11. Distribution Match
  12. Prior Reconstruction
  13. Post Reconstruction
  14. Implementation Details
  15. Experiments
  16. ...and 21 more sections

Key Result

Proposition 1

Maximizing the reconstruction term and/or minimizing the distribution matching term subsequently minimizes the EEA prediction matching term.

Figures (5)

  • Figure 1: Illustration of embedding-based entity alignment. The modules in the blue area belong to the EEA model, while those in the yellow area belong to the predictor.
  • Figure 2: The workflow of GEEA. Top: different sub-VAEs process different sub-embeddings, and the respective decoders convert the sub-embeddings back to concrete features. Bottom-left: the entity alignment prediction loss is retained. Bottom-center: the latent variables of sub-VAEs are used for distribution matching. Bottom-right: The reconstructed sub-embeddings are feed into the fusion layer in the EEA model to produce the reconstructed joint embedding for post reconstruction.
  • Figure 3: MRR results on FBDB15K, w.r.t. epochs.
  • Figure 4: Entity alignment results on FBDB15K, w.r.t. ratios of training alignment.
  • Figure 5: Entity alignment results on all datasets, w.r.t. ratios of training alignment.

Theorems & Definitions (5)

  • Proposition 1
  • Proposition 2
  • proof
  • proof
  • proof