Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Context-Aware Adapter Tuning for Few-Shot Relation Learning in Knowledge Graphs

Ran Liu, Zhongzhou Liu, Xiaoli Li, Yuan Fang

TL;DR

RelAdapter is a context-aware adapter for few-shot relation learning in KGs designed to enhance the adaptation process in meta-learning, and is validated over state-of-the-art methods on three benchmark KGs.

Abstract

Knowledge graphs (KGs) are instrumental in various real-world applications, yet they often suffer from incompleteness due to missing relations. To predict instances for novel relations with limited training examples, few-shot relation learning approaches have emerged, utilizing techniques such as meta-learning. However, the assumption is that novel relations in meta-testing and base relations in meta-training are independently and identically distributed, which may not hold in practice. To address the limitation, we propose RelAdapter, a context-aware adapter for few-shot relation learning in KGs designed to enhance the adaptation process in meta-learning. First, RelAdapter is equipped with a lightweight adapter module that facilitates relation-specific, tunable adaptation of meta-knowledge in a parameter-efficient manner. Second, RelAdapter is enriched with contextual information about the target relation, enabling enhanced adaptation to each distinct relation. Extensive experiments on three benchmark KGs validate the superiority of RelAdapter over state-of-the-art methods.

Context-Aware Adapter Tuning for Few-Shot Relation Learning in Knowledge Graphs

TL;DR

RelAdapter is a context-aware adapter for few-shot relation learning in KGs designed to enhance the adaptation process in meta-learning, and is validated over state-of-the-art methods on three benchmark KGs.

Abstract

Knowledge graphs (KGs) are instrumental in various real-world applications, yet they often suffer from incompleteness due to missing relations. To predict instances for novel relations with limited training examples, few-shot relation learning approaches have emerged, utilizing techniques such as meta-learning. However, the assumption is that novel relations in meta-testing and base relations in meta-training are independently and identically distributed, which may not hold in practice. To address the limitation, we propose RelAdapter, a context-aware adapter for few-shot relation learning in KGs designed to enhance the adaptation process in meta-learning. First, RelAdapter is equipped with a lightweight adapter module that facilitates relation-specific, tunable adaptation of meta-knowledge in a parameter-efficient manner. Second, RelAdapter is enriched with contextual information about the target relation, enabling enhanced adaptation to each distinct relation. Extensive experiments on three benchmark KGs validate the superiority of RelAdapter over state-of-the-art methods.

Paper Structure

This paper contains 18 sections, 6 equations, 7 figures, 7 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Related Work
  3. Preliminaries
  4. Methodology
  5. Context-aware Adapter
  6. Integration with Meta-learning
  7. Experiments
  8. Experiment Setup
  9. Comparison with Baselines
  10. Ablation Study
  11. Efficiency Analysis
  12. Sensitivity Analysis
  13. Conclusion
  14. Meta-training stage
  15. Datasets
  16. ...and 3 more sections

Figures (7)

  • Figure 1: Pairwise cosine similarity of relations.
  • Figure 2: Illustration of key concepts in RelAdapter, hinging on an entity-aware adapter (a, b) in the meta-testing stage (c). Note that we omit the meta-training stage, which is similar to meta-testing but with backpropagation of the query loss to update the model parameters ($\mathtt{emb}$ and $\Phi$).
  • Figure 3: Few-shot size
  • Figure 4: Adapter ratio
  • Figure 5: Context ratio
  • ...and 2 more figures