基于 ConceptFERE 的成像测井领域小样本关系抽取方法#br#

吉林大学学报(信息科学版) ›› 2026, Vol. 44 ›› Issue (3): 609-617.

基于 ConceptFERE 的成像测井领域小样本关系抽取方法#br#

曹茂俊¹, 焦俊齐¹, 李仲文¹, 吴润桐²

1. 东北石油大学计算机与信息技术学院,黑龙江大庆163318; 2. 大庆油田有限责任公司采油工艺研究院,黑龙江大庆163453

收稿日期:2025-05-28 出版日期:2026-06-02 发布日期:2026-06-02
通讯作者: 焦俊齐(2000— ), 男, 山东枣庄人, 东北石油大学硕士研究生,主要从事深度学习、人工智能研究,(Tel)86-18863269515(E-mail)jiaojunqi_01@126. com。 E-mail:jiaojunqi_01@126. com
作者简介:曹茂俊(1978— ),男,黑龙江大庆人,东北石油大学副教授,主要从事深度学习、智能计算在测井曲线识别中应用研究, (Tel)86-13796988520(E-mail)caomaojun@126. com
基金资助:
国家自然科学基金资助项目(42172161;52474035); 黑龙江省自然科学基金联合基金重点资助项目(ZL2024D003); 中石油创新基金资助项目(2024DQ02-0114)

Method for Few-Shot Relation Extraction in Imaging Logging Domain Based on ConceptFERE#br#

CAO Maojun¹, JIAO Junqi¹, LI Zhongwen¹, WU Runtong²

1. School of Computer and Information Technology, Northeast Petroleum University, Daqing 163318, China; 2. Oil Production Technology Research Institute, Daqing Oilfield Company Limited, Daqing 163453, China

Received:2025-05-28 Online:2026-06-02 Published:2026-06-02

摘要/Abstract

摘要： 针对现有成像测井领域数据缺乏、标注成本高, 以及传统关系抽取模型应用受限的问题, 提出了一种基于 ConceptFERE(Concept-Enhanced Few-Ehot Relation Extraction)模型的成像测井领域小样本关系抽取方法。首先, 基于 BERT-PAIR(BERT-Paired Sentence Encoding)框架, 研究了 ConceptFERE 模型的改进方法, 然后提出了改进模型 SDG-ConceptFERE(Semantic Difference Gate-ConceptFERE)。该模型通过引入语义差异门机制融合模块, 可动态判断支持集与查询集实例之间的相关性, 并仅对支持集实例融入外部实体概念信息, 从而有效避免了错误增强相关性导致的分类误差。实验结果表明, SDG-ConceptFERE 模型在5-way-1-shot和5-way-5-shot 任务设置下, 准确率与 ConceptFERE 模型相比分别提升3.57%和2.78%, 证明了其有效性。不仅为测井研究人员提供更方便的文本解释资料, 更有助于推动勘探开发全流程的智能化决策体系优化。

关键词: 成像测井, ConceptFERE模型, 小样本学习, 关系抽取, 语义差异门机制

Abstract: To address the challenges of data scarcity, high annotation costs, and limitations of traditional relation extraction models in imaging logging, a few-shot relation extraction method based on the ConceptFERE(Concept- Enhanced Few-Ehot Relation Extraction) model is proposed. Using the BERT-PAIR(BERT-Paired Sentence Encoding) framework, the ConceptFERE model is improved and the SDG-ConceptFERE(Semantic Difference Gate-ConceptFERE) model is introduced. It includes a semantic difference gate mechanism that dynamically assesses the relevance between support and query instances. By incorporating external entity concepts into the support instances, the classification errors from incorrect semantic enhancement are prevented. Experiments show SDG-ConceptFERE improves accuracy by 3.57% and 2.78% over ConceptFERE in 5-way-1-shot and 5-way-5- shot tasks, proving its effectiveness in providing better text support for logging researchers and advancing intelligent decision-making in exploration and development.

Key words: imaging logging, ConceptFERE model, few-shot learning, relation extraction, semantic differencegate mechanism

中图分类号:

TP391

曹茂俊, 焦俊齐, 李仲文, 吴润桐. 基于 ConceptFERE 的成像测井领域小样本关系抽取方法#br#[J]. 吉林大学学报(信息科学版), 2026, 44(3): 609-617.

CAO Maojun, JIAO Junqi, LI Zhongwen, WU Runtong. Method for Few-Shot Relation Extraction in Imaging Logging Domain Based on ConceptFERE#br#[J]. Journal of Jilin University (Information Science Edition), 2026, 44(3): 609-617.