Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

TransERR: Translation-based Knowledge Graph Embedding via Efficient Relation Rotation

Jiang Li, Xiangdong Su, Fujun Zhang, Guanglai Gao

TL;DR

Different from the previous translation-based models, TransERR encodes knowledge graphs in the hypercomplex-valued space, thus enabling it to possess a higher degree of translation freedom in mining latent information between the head and tail entities.

Abstract

This paper presents a translation-based knowledge geraph embedding method via efficient relation rotation (TransERR), a straightforward yet effective alternative to traditional translation-based knowledge graph embedding models. Different from the previous translation-based models, TransERR encodes knowledge graphs in the hypercomplex-valued space, thus enabling it to possess a higher degree of translation freedom in mining latent information between the head and tail entities. To further minimize the translation distance, TransERR adaptively rotates the head entity and the tail entity with their corresponding unit quaternions, which are learnable in model training. We also provide mathematical proofs to demonstrate the ability of TransERR in modeling various relation patterns, including symmetry, antisymmetry, inversion, composition, and subrelation patterns. The experiments on 10 benchmark datasets validate the effectiveness and the generalization of TransERR. The results also indicate that TransERR can better encode large-scale datasets with fewer parameters than the previous translation-based models. Our code and datasets are available at~\url{https://github.com/dellixx/TransERR}.

TransERR: Translation-based Knowledge Graph Embedding via Efficient Relation Rotation

TL;DR

Different from the previous translation-based models, TransERR encodes knowledge graphs in the hypercomplex-valued space, thus enabling it to possess a higher degree of translation freedom in mining latent information between the head and tail entities.

Abstract

This paper presents a translation-based knowledge geraph embedding method via efficient relation rotation (TransERR), a straightforward yet effective alternative to traditional translation-based knowledge graph embedding models. Different from the previous translation-based models, TransERR encodes knowledge graphs in the hypercomplex-valued space, thus enabling it to possess a higher degree of translation freedom in mining latent information between the head and tail entities. To further minimize the translation distance, TransERR adaptively rotates the head entity and the tail entity with their corresponding unit quaternions, which are learnable in model training. We also provide mathematical proofs to demonstrate the ability of TransERR in modeling various relation patterns, including symmetry, antisymmetry, inversion, composition, and subrelation patterns. The experiments on 10 benchmark datasets validate the effectiveness and the generalization of TransERR. The results also indicate that TransERR can better encode large-scale datasets with fewer parameters than the previous translation-based models. Our code and datasets are available at~\url{https://github.com/dellixx/TransERR}.
Paper Structure (18 sections, 6 equations, 3 figures, 9 tables)

This paper contains 18 sections, 6 equations, 3 figures, 9 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Background and Notation
  4. Methodology
  5. TransERR
  6. Theoretical Analysis
  7. Experimental Setup
  8. Datasets
  9. Evaluation Protocol
  10. Implementation
  11. Results and Analysis
  12. Main Results
  13. Complex Relations Analysis
  14. Key Relation Patterns Analysis
  15. Ablation Study
  16. ...and 3 more sections

Figures (3)

  • Figure 1: Illustration of TransE, RotatE and TransERR. TransE, RotatE and TransERR encode knowledge graphs in the real-valued space, complex-valued space and hypercomplex-valued space, respectively. $\circ$ denotes Hadamard product. The distance function of TransERR is $\parallel \mathbf{h} \otimes \mathbf{r^{\vartriangleleft H}} + \mathbf{r}- \mathbf{t} \otimes \mathbf{r^{\vartriangleleft T}} \parallel$.
  • Figure 2: Relation embedding histograms for various relation patterns. $\mathbf{r_1}$ is /music/performance_role/regular_performances. /music/group_membership/role. $\mathbf{r_2}$ and $\mathbf{r_3}$ are /film/actor/film./film/performance/film and /film/film/starring./film/performance/actor, respectively. $\mathbf{r_4}$, $\mathbf{r_5}$ and $\mathbf{r_6}$ are /people/person/nationality, /location/location/contains and /people/person/place_of_birth, respectively.
  • Figure 3: Visualisation of relation embeddings on Sports.