Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Empowering Few-Shot Relation Extraction with The Integration of Traditional RE Methods and Large Language Models

Ye Liu, Kai Zhang, Aoran Gan, Linan Yue, Feng Hu, Qi Liu, Enhong Chen

TL;DR

This work addresses the challenge of few-shot relation extraction by bridging traditional RE methods with Large Language Models. It introduces the Dual-System Augmented Relation Extractor (DSARE), which couples an LLM-augmented RE module, a RE-augmented LLM module, and an Integrated Prediction component to exploit complementary strengths and mitigate limitations of each approach. Through LLM-based data augmentation, KNN-driven demonstrations, and an integrated decision mechanism, DSARE achieves state-of-the-art performance on TACRED, TACREV, and Re-TACRED in few-shot regimes, with ablations underscoring the value of each component. The results suggest that jointly leveraging prior knowledge from LLMs and task-specific signals from traditional RE models can significantly improve few-shot relation extraction, offering a practical path for robust RE in low-resource settings.

Abstract

Few-Shot Relation Extraction (FSRE), a subtask of Relation Extraction (RE) that utilizes limited training instances, appeals to more researchers in Natural Language Processing (NLP) due to its capability to extract textual information in extremely low-resource scenarios. The primary methodologies employed for FSRE have been fine-tuning or prompt tuning techniques based on Pre-trained Language Models (PLMs). Recently, the emergence of Large Language Models (LLMs) has prompted numerous researchers to explore FSRE through In-Context Learning (ICL). However, there are substantial limitations associated with methods based on either traditional RE models or LLMs. Traditional RE models are hampered by a lack of necessary prior knowledge, while LLMs fall short in their task-specific capabilities for RE. To address these shortcomings, we propose a Dual-System Augmented Relation Extractor (DSARE), which synergistically combines traditional RE models with LLMs. Specifically, DSARE innovatively injects the prior knowledge of LLMs into traditional RE models, and conversely enhances LLMs' task-specific aptitude for RE through relation extraction augmentation. Moreover, an Integrated Prediction module is employed to jointly consider these two respective predictions and derive the final results. Extensive experiments demonstrate the efficacy of our proposed method.

Empowering Few-Shot Relation Extraction with The Integration of Traditional RE Methods and Large Language Models

TL;DR

This work addresses the challenge of few-shot relation extraction by bridging traditional RE methods with Large Language Models. It introduces the Dual-System Augmented Relation Extractor (DSARE), which couples an LLM-augmented RE module, a RE-augmented LLM module, and an Integrated Prediction component to exploit complementary strengths and mitigate limitations of each approach. Through LLM-based data augmentation, KNN-driven demonstrations, and an integrated decision mechanism, DSARE achieves state-of-the-art performance on TACRED, TACREV, and Re-TACRED in few-shot regimes, with ablations underscoring the value of each component. The results suggest that jointly leveraging prior knowledge from LLMs and task-specific signals from traditional RE models can significantly improve few-shot relation extraction, offering a practical path for robust RE in low-resource settings.

Abstract

Few-Shot Relation Extraction (FSRE), a subtask of Relation Extraction (RE) that utilizes limited training instances, appeals to more researchers in Natural Language Processing (NLP) due to its capability to extract textual information in extremely low-resource scenarios. The primary methodologies employed for FSRE have been fine-tuning or prompt tuning techniques based on Pre-trained Language Models (PLMs). Recently, the emergence of Large Language Models (LLMs) has prompted numerous researchers to explore FSRE through In-Context Learning (ICL). However, there are substantial limitations associated with methods based on either traditional RE models or LLMs. Traditional RE models are hampered by a lack of necessary prior knowledge, while LLMs fall short in their task-specific capabilities for RE. To address these shortcomings, we propose a Dual-System Augmented Relation Extractor (DSARE), which synergistically combines traditional RE models with LLMs. Specifically, DSARE innovatively injects the prior knowledge of LLMs into traditional RE models, and conversely enhances LLMs' task-specific aptitude for RE through relation extraction augmentation. Moreover, an Integrated Prediction module is employed to jointly consider these two respective predictions and derive the final results. Extensive experiments demonstrate the efficacy of our proposed method.
Paper Structure (23 sections, 3 figures, 3 tables)

This paper contains 23 sections, 3 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Few-shot Relation Extraction.
  4. Large Language Models.
  5. Problem Statement
  6. DSARE Model
  7. LLM-augmented RE
  8. LLM Data Augmentation.
  9. Traditional RE Model.
  10. RE-augmented LLM
  11. KNN Demonstration.
  12. LLM Inference.
  13. Integrated Prediction
  14. Experiments
  15. Experimental Setup
  16. ...and 8 more sections

Figures (3)

  • Figure 1: The comparison between traditional RE methods and LLM-based RE methods.
  • Figure 2: The architecture of DSARE. It includes three parts: (a) LLM-augmented RE module; (b) RE-augmented LLM module; (c) Integrated Prediction module.
  • Figure 3: The case study of the DSARE model. (a) is from the TACRED dataset (K=8), while (b) and (c) are from the Re-TACRED dataset (K=8).