融合分段式位置信息的图卷积中文关系抽取

doi:10.3778/j.issn.1002-8331.2404-0246

摘要/Abstract

摘要： 关系抽取任务旨在识别句子中实体间的语义关系。针对中文关系抽取任务中语义信息挖掘不充分以及引入语法依赖产生的噪声问题，提出一种融合分段式位置信息的图卷积中文关系抽取模型。一方面引入实体的位置信息，在初始文本上添加相对位置编码，通过分段卷积神经网络将实体位置信息集成到局部特征中，并引入门控注意力机制构建注意力权重矩阵，捕获全局语义特征。另一方面通过语法剪枝规则构建局部语法依赖图捕获语法信息，去除与特定实体语法距离较远且无关的分支来避免噪声干扰。最后采用门控机制动态融合不同粒度的语义信息，实现不同类型信息的共享与互补。在三个公开中文数据集SanWen、FinRE和COAE2016上的实验结果表明，该模型可以有效捕获语义语法信息，对比基线模型表现出了较好的性能。

关键词: 中文关系抽取, 图卷积网络, 分段式位置信息, 语法剪枝, 门控注意力

Abstract: The relationship extraction task aims to recognize the semantic relationships between entities in a sentence. Aiming at the problems of insufficient semantic information mining and the noise generated by introducing syntactic dependencies in the Chinese relation extraction task, a graph convolutional Chinese relation extraction model incorporating piecewise position information is proposed. On the one hand, the position information of entities is introduced, the relative position encoding is added to the initial text, the entity position information is integrated into the local features by segmented convolutional neural network, and the gated attention mechanism is introduced to construct the attention weight matrix to capture the global semantic features. On the other hand, the local syntactic dependency graph is constructed by syntactic pruning rule to capture the syntactic information, and the branches that are far away and irrelevant to the syntax of a specific entity are removed to avoid noise interference. Finally, a gated mechanism is used to dynamically fuse semantic information of different granularities to realize the sharing and complementarity of different types of information. Experimental results on three publicly available Chinese datasets, SanWen, FinRE, and COAE2016, show that the model can effectively capture semantic syntactic information and exhibits better performance than the baseline model.

Key words: Chinese relation extraction, graph convolutional network, piecewise location information, syntactic pruning, gated attention

王婷婷, 韩虎, 何勇禧, 孔博. 融合分段式位置信息的图卷积中文关系抽取[J]. 计算机工程与应用, 2025, 61(16): 187-195.

WANG Tingting, HAN Hu, HE Yongxi, KONG Bo. Graph Convolutional Chinese Relation Extraction Fused with Piecewise Position Information[J]. Computer Engineering and Applications, 2025, 61(16): 187-195.

参考文献

[1] TRISEDYA B D, WEIKUM G, QI J Z, et al. Neural relation extraction for knowledge base enrichment[C]//Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics. Stroudsburg: ACL, 2019: 229-240.
[2] DENG Y, XIE Y X, LI Y L, et al. Multi-task learning with multi-view attention for answer selection and knowledge base question answering[J]. Proceedings of the AAAI Conference on Artificial Intelligence, 2019, 33(1): 6318-6325.
[3] ZHOU M, DUAN N, LIU S J, et al. Progress in neural NLP: modeling, learning, and reasoning[J]. Engineering, 2020, 6(3): 155-188.
[4] WANG X, HE X N, CAO Y X, et al. Knowledge graph attention network for recommendation[C]//Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. New York: ACM, 2019: 950-958.
[5] ZENG D J, LIU K, CHEN Y B, et al. Distant supervision for relation extraction via piecewise convolutional neural networks[C]//Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing. Stroudsburg: ACL, 2015: 1753-1762.
[6] 隗昊, 唐焕玲, 周爱, 等. 基于双路分段注意力神经张量网络的临床文本关系抽取[J]. 电子学报, 2023, 51(3): 658-665.
KUI H, TANG H L, ZHOU A, et al. Clinical relation extraction via dual piecewise attention neural tensor network[J]. Acta Electronica Sinica, 2023, 51(3): 658-665.
[7] 闫雄, 段跃兴, 张泽华. 采用自注意力机制和CNN融合的实体关系抽取[J]. 计算机工程与科学, 2020, 42(11): 2059-2066.
YAN X, DUAN Y X, ZHANG Z H. Entity relationship extraction fusing self-attention mechanism and CNN[J]. Computer Engineering & Science, 2020, 42(11): 2059-2066.
[8] SUN C H, JI W D, ZHOU G H, et al. FGSI: distant supervision for relation extraction method based on fine-grained semantic information[J]. Scientific Reports, 2023, 13: 14075.
[9] 肖立中, 臧中兴, 宋赛赛. 融合自注意力的关系抽取级联标记框架研究[J]. 计算机工程与应用, 2023, 59(3): 77-83.
XIAO L Z, ZANG Z X, SONG S S. Research on cascaded labeling framework for relation extraction with self-attention[J]. Computer Engineering and Applications, 2023, 59(3): 77-83.
[10] LONG J, WANG Y, WEI X X, et al. Entity-centric fully connected GCN for relation classification[J]. Applied Sciences, 2021, 11(4): 1377.
[11] LI D, LEI Z L, SONG B Y, et al. Neural attentional relation extraction with dual dependency trees[J]. Journal of Computer Science and Technology, 2022, 37(6): 1369-1381.
[12] CHEN X Y, ZHANG M, XIONG S W, et al. On the form of parsed sentences for relation extraction[J]. Knowledge-Based Systems, 2022, 251: 109184.
[13] XU Y, MOU L L, LI G, et al. Classifying relations via long short term memory networks along shortest dependency paths[C]//Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing. Stroudsburg: ACL, 2015: 1785-1794.
[14] LI J J, SHUANG K, GUO J Y, et al. Enhancing semantic relation classification with shortest dependency path reasoning[J]. IEEE/ACM Transactions on Audio, Speech, and Language Processing, 2023, 31: 1550-1560.
[15] 李冬梅, 张扬, 李东远, 等. 实体关系抽取方法研究综述[J]. 计算机研究与发展, 2020, 57(7): 1424-1448.
LI D M, ZHANG Y, LI D Y, et al. Review of entity relation extraction methods[J]. Journal of Computer Research and Development, 2020, 57(7): 1424-1448.
[16] 刘成星, 张超群, 代林林, 等. TFLS-BiGRU-ATT: 一种简单有效的中文短文本关系抽取模型[J]. 中文信息学报, 2023, 37(6): 115-127.
LIU C X, ZHANG C Q, DAI L L, et al. TFLS-BiGRU-ATT: a simple but effective relation extraction model for Chinese short text[J]. Journal of Chinese Information Processing, 2023, 37(6): 115-127.
[17] ZENG D J, LIU K, LAI S W, et al. Relation classification via convolutional deep neural network[C]//Proceedings of the 25th International Conference on Computational Linguistics: Technical Papers. Stroudsburg: ACL, 2014, 2335-2344.
[18] ZHANG D X, WANG D. Relation classification via recurrent neural network[J]. arXiv:1508.01006, 2015.
[19] WELLING M, KIPF T N. Semi-supervised classification with graph convolutional networks[C]//Proceedings of the 4th International Conference on Learning Representations, 2016: 1-6.
[20] QIAN Y J, SANTUS E, JIN Z J, et al. GraphIE: a graph-based framework for information extraction[C]//Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics. Stroudsburg: ACL, 2019: 751-761.
[21] DEVLIN J, CHANG M W, LEE K, et al. BERT: pretraining of deep bidirectional transformers for language understanding[C]//Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics. Stroudsburg: ACL, 2019: 4171-4186.
[22] HUA W Z, DAI Z H, LIU H X, et al. Transformer quality in linear time[C]//Proceedings of the 39th International Conference on Machine Learning, 2022: 9099-9117.
[23] XU J J, WEI J, XU S, et al. A discourse-level named entity recognition and relation extraction dataset for Chinese literature text[J]. arXiv:1711.07010, 2017.
[24] LI Z R, DING N, LIU Z Y, et al. Chinese relation extraction with multi-grained information and external linguistic knowledge[C]//Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics. Stroudsburg: ACL, 2019: 4377-4386.
[25] LE H Q, CAN D C, VU S T, et al. Large-scale exploration of neural relation classification architectures[C]//Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing. Stroudsburg: ACL, 2018: 2266-2277.
[26] ZHONG Z C. Chinese entity relation extraction based on multi-level gated recurrent mechanism and self-attention[C]//Proceedings of the 2021 2nd International Conference on Artificial Intelligence and Information Systems. New York: ACM, 2021: 1-7.
[27] ZHAO Q H, GAO T H, GUO N. A novel Chinese relation extraction method using polysemy rethinking mechanism[J]. Applied Intelligence, 2023, 53(7): 7665-7676.
[28] YANG J, JI B, LI S S, et al. Dynamic multi-view fusion mechanism for Chinese relation extraction[C]//Proceedings of the 27th Pacific-Asia Conference on Knowledge Discovery and Data Mining. Cham: Springer, 2023: 405-417.
[29] 孟佳营. 基于句法分析的关系分类和关系抽取方法研究[D]. 杭州: 杭州电子科技大学, 2022.
MENG J Y. Research on relation classification and relation extraction based on syntactic analysis[D]. Hangzhou: Hangzhou Dianzi University, 2022.
[30] XU B, LI S, ZHANG Z W, et al. BERT-PAGG: a Chinese relationship extraction model fusing PAGG and entity location information[J]. PeerJ Computer Science, 2023, 9: e1470.