融合异构图网络的多轮对话实体关系抽取

doi:10.3778/j.issn.1002-8331.2401-0321

摘要/Abstract

摘要： 对话实体关系抽取旨在从对话中提取实体对之间的关系。在多轮对话的情境下，实体信息可能分散在各个对话轮次中，同时还可能受到语境变化和语言表达多样性的影响，从而使多轮对话中的实体关系抽取更具挑战性。提出了一种融合异构图网络的多轮对话实体关系抽取模型。该模型通过文本解释器获取上下文表示，以捕捉每轮对话中的实体信息及其相应的语境。利用高斯多视图模块提取文本的隐含特征，捕捉多种视角下的对话信息。为了融合多轮对话中分散的关系信息，引入了异构图的概念，并利用异构图卷积进行信息传递。最后通过分类器对实体对之间的关系进行识别。在公共数据集DialogRE上的实验证明了该模型的有效性，其F1值为73.9%，F1c值为67.4%。

关键词: 关系抽取, 多轮对话, 异构图, 多视图

Abstract: The extraction of entity relation in dialogue aims to extract relationships between entities within a dialogue. In the context of multi-turn dialogue, entity information may be dispersed across various dialogue turns, subject to contextual variations and linguistic diversity, thereby extracting entity relations in multi-turn dialogues becomes more challenging. A multi-turn dialogue entity relation extraction model of heterogeneous graph networks is proposed. The model employs a text interpreter to obtain contextual representations, capturing entity information and its corresponding context in each dialogue turn. The Gaussian multi-view module is utilized to extract hidden features from the text, capturing dialogue information from multiple perspectives. To integrate dispersed relationship information in multi-turn dialogues, the concept of a heterogeneous graph is introduced, and heterogeneous graph convolution is utilized for information propagation. Finally, a classifier is utilized to identify relationships among pairs of entities. The model has been proven valid by experiment results on the public dataset DialogRE, with an F1 score of 73.9% and an F1c score of 67.4%.

Key words: relation extraction, multi-turn dialogue, heterogeneous graph, multi-view

张顺淼, 郑思源. 融合异构图网络的多轮对话实体关系抽取[J]. 计算机工程与应用, 2025, 61(10): 176-184.

ZHANG Shunmiao, ZHENG Siyuan. Multi-Turn Dialogue Entity Relation Extraction of Heterogeneous Graph Networks[J]. Computer Engineering and Applications, 2025, 61(10): 176-184.

参考文献

[1] 陆亮, 孔芳. 面向对话的融入交互信息的实体关系抽取[J]. 中文信息学报, 2021, 35(8): 82-88.
LU L, KONG F. Interactive information enhanced relation extraction from dialogue[J]. Journal of Chinese Information Processing, 2021, 35(8): 82-88.
[2] 陆亮, 孔芳. 面向对话的融入知识的实体关系抽取[J]. 计算机科学, 2022, 49(5): 200-205.
LU L, KONG F. Dialogue-based entity relation extraction with knowledge[J]. Computer Science, 2022, 49(5): 200-205.
[3] YU D, SUN K, CARDIE C, et al. Dialogue-based relation extraction[C]//Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics. Stroudsburg: ACL, 2020: 4927-4940.
[4] DEVLIN J, CH M, Lee K, et al. BERT: pre-training of deep bidirectional transformers for language understanding[C]//Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies. Stroudsburg: ACL, 2019: 4171-4186.
[5] 马建红, 龚天, 姚爽. 基于证据句与图卷积网络的文档级关系抽取[J]. 计算机工程, 2023, 49(8): 104-110.
MA J H, GONG T, YAO S. Document-level relation extraction based on evidential sentences and graph convolutional network[J]. Computer Engineering, 2023, 49(8): 104-110.
[6] ZENG D, LIU K, LAI S, et al. Relation classification via convolutional deep neural network[C]//Proceedings of the 25th International Conference on Computational Linguistics, 2014: 2335-2344.
[7] ZHANG S, ZHENG D, HU X, et al. Bidirectional long short term memory networks for relation classification[C]//Proceedings of the 29th Pacific Asia Conference on Language. New York: ACM, 2015: 73-78.
[8] LIU C Y, SUN W B, CHAO W H, et al. Convolution neural network for relation extraction[M]//Advanced data mining and applications. Berlin, Heidelberg: Springer, 2013: 231-242.
[9] NGUYEN T H, GRISHMAN R. Relation extraction: perspective from convolutional neural networks[C]//Proceedings of the 1st Workshop on Vector Space Modeling for Natural Language Processing. Stroudsburg: ACL, 2015: 39-48.
[10] XUE F Z, SUN A X, ZHANG H, et al. An embarrassingly simple model for dialogue relation extraction[C]//Proceedings of the 2022 IEEE International Conference on Acoustics, Speech and Signal Processing. Piscataway: IEEE, 2022: 6707-6711.
[11] VU N T, ADEL H, GUPTA P, et al. Combining recurrent and convolutional neural networks for relation classification[C]//Proceedings of the 2016 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies. Stroudsburg: ACL, 2016: 534-539.
[12] ZHANG Y H, ZHONG V, CHEN D Q, et al. Position-aware attention and supervised data improve slot filling[C]//Proceedings of the 2017 Conference on Empirical Methods in Natural Language Processing. Stroudsburg: ACL, 2017: 35-45.
[13] KATIYAR A, CARDIE C. Going out on a limb: joint extraction of entity mentions and relations without dependency trees[C]//Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics. Stroudsburg: ACL, 2017: 917-928.
[14] 唐媛, 陈艳平, 扈应, 等. 一种面向关系抽取的表填充依赖特征学习方法[J]. 计算机工程与应用, 2024, 60(13): 143-151.
TANG Y, CHEN Y P, HU Y, et al. Dependency feature learning method for table filling for relation extraction[J]. Computer Engineering and Applications, 2024, 60(13): 143-151.
[15] SON J, KIM J, LI J, et al. GRASP: guiding model with relational semantics using prompt for dialogue relation extraction[C]//Proceedings of the 29th International Conference on Computational Linguistics, 2022: 412-423.
[16] XU W, CHEN K H, ZHAO T J. Document-level relation extraction with reconstruction[J]. Proceedings of the AAAI Conference on Artificial Intelligence, 2021, 35(16): 14167-14175.
[17] ZHANG Y H, QI P, MANNING C D. Graph convolution over pruned dependency trees improves relation extraction[C]//Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing. Stroudsburg: ACL, 2018: 2205-2215.
[18] GUO Z J, ZHANG Y, LU W. Attention guided graph convolutional networks for relation extraction[C]//Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics. Stroudsburg: ACL, 2019: 241-251.
[19] XUE F Z, SUN A X, ZHANG H, et al. GDPNet: refining latent multi-view graph for relation extraction[J]. Proceedings of the AAAI Conference on Artificial Intelligence, 2021, 35(16): 14194-14202.
[20] GHOSAL D, MAJUMDER N, PORIA S, et al. DialogueGCN: a graph convolutional neural network for emotion recognition in conversation[C]//Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing. Stroudsburg: ACL, 2019: 154-164.
[21] CHEN H, HONG P F, HAN W, et al. Dialogue relation extraction with document-level heterogeneous graph attention networks[J]. Cognitive Computation, 2023, 15(2): 793-802.
[22] ISHIWATARI T, YASUDA Y, MIYAZAKI T, et al. Relation-aware graph attention networks with relational position encodings for emotion recognition in conversations[C]//Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing. Stroudsburg: ACL, 2020: 7360-7370.
[23] LEE B, CHOI Y S. Graph based network with contextualized representations of turns in dialogue[C]//Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing. Stroudsburg: ACL, 2021: 443-455.
[24] LIU X, ZHANG J, ZHANG H, et al. Hierarchical dialogue understanding with special tokens and turn-level attention[J]. arXiv:2305.00262, 2023.
[25] 王琪琪, 李培峰. 基于GCN的多人对话实体关系抽取方法[J]. 中文信息学报, 2023, 37(5): 80-87.
WANG Q Q, LI P F. A GCN-based approach to entity relation extraction from multi-party dialogues[J]. Journal of Chinese Information Processing, 2023, 37(5): 80-87.
[26] 自彦丞, 李卫疆. TDGCN: 触发器增强的两阶段动态图卷积网络的对话关系抽取研究[J]. 小型微型计算机系统, 2025, 46(1): 90-96.
ZI Y C, LI W J. TDGCN: research on conversation relationship extraction of two stage dynamic graph convolutional networks enhanced by triggers[J]. Journal of Chinese Computer Systems, 2025, 46(1): 90-96.
[27] 徐洋, 蒋玉茹, 张禹尧, 等. 融合角色指代的多方对话关系抽取方法研究[J]. 北京大学学报 (自然科学版), 2022, 58(1): 13-20.
XU Y, JIANG Y R, ZHANG Y Y, et al. Research on dialogue entity relation extraction with enhancing character information[J]. Acta Scientiarum Naturalium Universitatis Pekinensis, 2022, 58(1): 13-20.
[28] FEI H, LI J Y, WU S Q, et al. Global inference with explicit syntactic and discourse structures for dialogue-level relation extraction[C]//Proceedings of the 31st International Joint Conference on Artificial Intelligence, 2022: 4107-4113.
[29] LIN P W, SU S Y, CHEN Y N. TREND: trigger-enhanced relation-extraction network for dialogues[C]//Proceedings of the 23rd Annual Meeting of the Special Interest Group on Discourse and Dialogue, 2022: 623-629.
[30] AN H, CHEN D S, XU W Y, et al. TLAG: an informative trigger and label-aware knowledge guided model for dialogue-based relation extraction[C]//Proceedings of the 2023 26th International Conference on Computer Supported Cooperative Work in Design, 2023: 59-64.
[31] ZHOU Y, LEE W S. None class ranking loss for document-level relation extraction[C]//Proceedings of the 31st International Joint Conference on Artificial Intelligence, 2022: 4538-4544.
[32] DUAN G D, DONG Y R, MIAO J Y, et al. Position-aware attention mechanism-based bi-graph for dialogue relation extraction[J]. Cognitive Computation, 2023, 15(1): 359-372.
[33] BAI X, SONG Y. Semantic-based pre-training for dialogue understanding[C]//Proceedings of the 29th International Conference on Computational Linguistics, 2022: 592-607.
[34] LIU Y, OTT M, GOYAL N et al. RoBERTa: a robustly optimized BERT pretraining approach[J] arXiv:1907.11692, 2019.