Lune

EMNLP Findings2024

Block-Diagonal Orthogonal Relation and Matrix Entity for Knowledge Graph Embedding

Yihua Zhu, Hidetoshi Shimodaira

Abstract

The primary aim of Knowledge Graph Embeddings (KGE) is to learn low-dimensional representations of entities and relations for predicting missing facts. Although rotation-based methods like RotatE (Sun et al., 2019) and QuatE (Zhang et al., 2019) perform well in KGE, they face two challenges: limited model flexibility requiring proportional increases in relation size with entity dimension, and difficulties in generalizing the model for higherdimensional rotations. To address these issues, we introduce OrthogonalE, a novel KGE model employing matrices for entities and block-diagonal orthogonal matrices with Riemannian optimization for relations. This approach not only enhances the generality and flexibility of KGE models but also captures several relation patterns that rotation-based methods can identify. Experimental results indicate that our new KGE model, OrthogonalE, offers generality and flexibility, captures several relation patterns, and significantly outperforms state-of-the-art KGE models while substantially reducing the number of relation parameters.

Ask Lune

Know this paper in minutes, not an afternoon.

Lune can read every page of it and the top-tier papers that cite it, and gives you its claims, its evidence and its limits, each one cited to the paper it comes from. No skimming, no AI slop.

Questions to start from

Lune can draw on

the full paper · related work

Get the answer

Free to start. No credit card required.

Lineage

  • HittER: Hierarchical Transformers for Knowledge Graph Embeddings cites Orthogonal Relation Transforms with Graph Context Modeling for Knowledge Graph Embedding
Builds onCited byLeft to right by year; the more cited, the higher.

Builds on6

  1. HittER: Hierarchical Transformers for Knowledge Graph EmbeddingsSanxing Chen, Xiaodong Liu, Jianfeng Gao, Jian Jiao et al.EMNLP 2021 · 110 citations
  2. Orthogonal Relation Transforms with Graph Context Modeling for Knowledge Graph EmbeddingYun Tang, Jing Huang, Guangtao Wang, Xiaodong He et al.ACL 2020 · 92 citations
  3. Unsupervised Learning of Equivariant Structure from SequencesTakeru Miyato, Masanori Koyama, Kenji FukumizuNeurIPS 2022 · 17 citations
  4. Coordinate Descent on the Orthogonal Group for Recurrent Neural Network TrainingEstelle M. Massart, Vinayak AbrolAAAI 2022 · 13 citations
  5. Generalizing Knowledge Graph Embedding with Universal Orthogonal ParameterizationRui Li, Chaozhuo Li, Yanming Shen, Zeyu Zhang et al.ICML 2024 · 11 citations

Related papers

  1. HousE: Knowledge Graph Embedding with Householder ParameterizationRui Li, Jianan Zhao, Chaozhuo Li, Di He et al.ICML 2022 · 66 citations
  2. Dual Quaternion Knowledge Graph EmbeddingsZongsheng Cao, Qianqian Xu, Zhiyong Yang, Xiaochun Cao et al.AAAI 2021 · 186 citations
  3. Improving Knowledge Graph Embedding Using Affine Transformations of Entities Corresponding to Each RelationJinfa Yang, Yongjie Shi, Xin Tong, Robin Wang et al.EMNLP Findings 2021 · 5 citations
  4. Enhancing Hyperbolic Knowledge Graph Embeddings via Lorentz TransformationsXiran Fan, Minghua Xu, Huiyuan Chen, Yuzhong Chen et al.ACL Findings 2024 · 2 citations
  5. BiQUE: Biquaternionic Embeddings of Knowledge GraphsJia Guo, Stanley KokEMNLP 2021
Dusk over the sea between two cliffs drawn in fine vertical lines