计算机科学与探索2024，Vol.18Issue(9)：2326-2336,11.DOI:10.3778/j.issn.1673-9418.2406056
基于大语言模型增强表征对齐的小样本持续关系抽取方法
Large Language Model Augmentation and Feature Alignment Method for Few-Shot Continual Relation Extraction
摘要
Abstract
Relation extraction,as a key task in natural language processing,plays a significant role in deepening lan-guage understanding,constructing knowledge graphs,and optimizing information retrieval systems.However,tradi-tional supervised learning methods are not well-suited for real-world scenarios due to the continuous emergence of new relations and the lack of large annotated datasets.Although the advent of large language models has significantly improved the performance of many natural language processing tasks,they still cannot effectively address the chal-lenges of few-shot continual relation extraction.To fully leverage the semantic knowledge of large language models to mitigate catastrophic forgetting and overfitting issues,a novel few-shot continual relation extraction method,LAFA(large language model augmentation and feature alignment),is proposed.This method enhances representation align-ment through various strategies such as relation instance rewriting,semantic expansion,and enhanced relation repre-sentation.It effectively improves the model adaptability to new relations and the retention of old knowledge while maintaining low data and computational costs.Experimental validation on two relation extraction datasets,FewRel and TACRED,demonstrates that LAFA outperforms existing methods in few-shot continual relation extraction tasks,particularly achieving the best results in incremental stages.Ablation experiments further reveal the signifi-cant contributions of each module to overall performance.Moreover,the inference efficiency and cost of LAFA are substantially lower than those of existing large language model-based methods,and it boasts strong scalability,being able to adapt to various language models.关键词
大语言模型(LLM)/关系抽取/持续学习/小样本学习Key words
large language model(LLM)/relation extraction/continual learning/few-shot learning分类
信息技术与安全科学引用本文复制引用
李逸飞,张玲玲,董宇轩,王佳欣,仲宇杰,魏笔凡..基于大语言模型增强表征对齐的小样本持续关系抽取方法[J].计算机科学与探索,2024,18(9):2326-2336,11.基金项目
国家重点研发计划(2022YFC3303600) （2022YFC3303600）
国家自然科学基金(62137002,62293553,62293554,62176207,62106190,62192781) （62137002,62293553,62293554,62176207,62106190,62192781）
传播内容认知全国重点实验室科研课题资助项目(A202402) （A202402）
陕西省自然科学基金(2023-JC-YB-593) （2023-JC-YB-593）
陕西高校青年创新团队项目. This work was supported by the National Key Research and Development Program of China(2022YFC3303600),the National Natural Science Foundation of China(62137002,62293553,62293554,62176207,62106190,62192781),the Research Project of State Key Laboratory of Communication Content Cognition(A202402),the Natural Science Foundation of Shaanxi Province(2023-JC-YB-593),and the Youth Innovation Team Project of Shaanxi Universities. （2022YFC3303600）
