结合二维增强融合机制的事件论元抽取方法

doi:10.3778/j.issn.1002-8331.2402-0185

摘要/Abstract

摘要： 针对现有的事件论元抽取研究中触发词和论元间缺少交互以及通道内部缺少交互的问题，提出结合二维增强融合机制的事件论元抽取模型（two-dimensional enhanced fusion mechanism for event argument extraction，W2-ARG）。在句子中的触发词两边插入标识符，引入事件类型信息的同时增强触发词和论元的交互，并单独编码触发词以突出其在句子中的语义信息。将论元抽取建模为二维化表示的标签预测，通过膨胀卷积捕获不同距离的单词的语义交互。使用通道注意力模块增强通道内部的交互，以强化通道内的信息传递。最后利用拉普拉斯算子来突出事件论元在语义空间中的位置特征，提升模型对事件论元的识别精度。模型在ACE05-EN、ERE-EN数据集上进行了实验，实验结果表明该方法的性能相较其他基于分类的事件论元抽取方法提升明显。

关键词: 事件论元抽取, 句子平面化表示, 通道注意力, 拉普拉斯算子, BERT

Abstract: Addressing the lack of interaction between trigger and arguments, and the lack of interaction within channels in existing studies of event argument extraction, this paper proposes a two-dimensional enhanced fusion mechanism for event argument extraction (W2-ARG) that incorporates a 2D enhanced fusion method. It inserts identifiers on both sides of a trigger word in a sentence, which encodes information about the event type. In the learning process, it is effective to enhance the interaction between trigger and argument, and encode the trigger separately to highlight its information in the sentence. The event argument extraction is implemented as a two-dimensional representation of label prediction. It has the advantage to capture semantic interactions of words at different distances through dilated convolution. Subsequently, a channel attention module is applied to enhance the interactions within the channel to strengthen the information transfer within the channel. Finally, the Laplace operator is utilized to highlight the positional features of event arguments in the semantic space to improve the performance. The proposed model is experimented on the ACE05-EN and ERE-EN datasets. Experimental results show clear performance improvement of the proposed model compared with other event argument extraction methods.

Key words: event argument extraction, sentence planarization representation, channel attention, Laplace operator, BERT

王潞翔, 陈艳平, 黄辉, 黄瑞章, 秦永彬. 结合二维增强融合机制的事件论元抽取方法[J]. 计算机工程与应用, 2025, 61(10): 111-119.

WANG Luxiang, CHEN Yanping, HUANG Hui, HUANG Ruizhang, QIN Yongbin. Event Argument Extraction with Two-Dimensional Enhanced Fusion Mechanism[J]. Computer Engineering and Applications, 2025, 61(10): 111-119.

参考文献

[1] DODDINGTON G, MITCHELL A, PRZYBOCKI M A, et al. The automatic content extraction (ACE) program-tasks, data, and evaluation[C]//Proceedings of the 2004 International Conference on Language Resources and Evaluation, 2004.
[2] 胡杭乐, 程春雷, 叶青, 等. 开放信息抽取研究综述[J]. 计算机工程与应用, 2023, 59(16): 31-49.
HU H L, CHENG C L, YE Q, et al. Survey of open information extraction research[J]. Computer Engineering and Applications, 2023, 59(16): 31-49.
[3] SOUZA COSTA T, GOTTSCHALK S, DEMIDOVA E, et al. Event-QA[C]//Proceedings of the 29th ACM International Conference on Information & Knowledge Management. New York: ACM, 2020: 3157-3164.
[4] BOSSELUT A, LE BRAS R, CHOI Y. Dynamic neuro-symbolic knowledge graph construction for zero-shot commonsense question answering[J]. Proceedings of the AAAI Conference on Artificial Intelligence, 2021, 35(6): 4923-4931.
[5] CHEN Y B, XU L H, LIU K, et al. Event extraction via dynamic multi-pooling convolutional neural networks[C]//Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Conference on Natural Language Processing. Stroudsburg: ACL, 2015: 167-176.
[6] HU C, LI J Y, SU F F, et al. OneEE: a one-stage framework for fast overlapping and nested event extraction[C]//Proceedings of the 29th International Conference on Computational Linguistics, 2022: 1953-1964.
[7] LI J Y, FEI H, LIU J, et al. Unified named entity recognition as word-word relation classification[J]. Proceedings of the AAAI Conference on Artificial Intelligence, 2022, 36(10): 10965-10973.
[8] DEVLIN J, CHANG M W, 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.
[9] SCHUSTER M, PALIWAL K K. Bidirectional recurrent neural networks[J]. IEEE Transactions on Signal Processing, 1997, 45(11): 2673-2681.
[10] WANG Q L, WU B G, ZHU P F, et al. ECA-net: efficient channel attention for deep convolutional neural networks[C]//Proceedings of the 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2020: 11531-11539.
[11] YU J T, BOHNET B, POESIO M. Named entity recognition as dependency parsing[C]//Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics. Stroudsburg: ACL, 2020: 6470-6476.
[12] SHA L, QIAN F, CHANG B B, et al. Jointly extracting event triggers and arguments by dependency-bridge RNN and tensor-based argument interaction[J]. Proceedings of the AAAI Conference on Artificial Intelligence, 2018, 32(1): 5916-5923.
[13] QIU X P, SUN T X, XU Y G, et al. Pre-trained models for natural language processing: a survey[J]. Science China Technological Sciences, 2020, 63(10): 1872-1897.
[14] RADFORD A, NARASIMHAN K, SALIMANS T, et al. Improving language understanding by generative pretraining[EB/OL]. (2018) [2024-06-30]. https://www.semanticscholar.org/paper/Improving-Language-Understanding-by-Generative-Radford?Narasimhan/cd18800a0fe0b668a1cc19f2ec95b
5003d0a5035.
[15] PETERS M, NEUMANN M, IYYER M, et al. Deep contextualized word representations[C]//Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human LanguageTechnologies. Stroudsburg: ACL, 2018: 2227-2237.
[16] WADDEN D, WENNBERG U, LUAN Y, et al. Entity, relation, and event extraction with contextualized span representations[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: 5784-5789.
[17] DU X Y, CARDIE C. Event extraction by answering (almost) natural questions[C]//Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing. Stroudsburg: ACL, 2020: 671-683.
[18] LU Y J, LIN H Y, XU J, et al. Text2Event: controllable sequence-to-structure generation for end-to-end event extraction[C]//Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing. Stroudsburg: ACL, 2021: 2795-2806.
[19] HAN R D, PENG T, YANG C H, et al. Is information extraction solved by ChatGPT? An analysis of performance, evaluation criteria, robustness and errors[J]. arXiv:2305. 14450, 2023.
[20] LIN Y, JI H, HUANG F, et al. A joint neural model for information extraction with global features[C]//Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics. Stroudsburg: ACL, 2020: 7999-8009.
[21] WANG S J, YU M, CHANG S Y, et al. Query and extract: refining event extraction as type-oriented binary decoding[C]//Findings of the Association for Computational Linguistics: ACL 2022. Stroudsburg: ACL, 2022: 169-182.
[22] ZHANG Z X, JI H. Abstract meaning representation guided graph encoding and decoding for joint information extraction[C]//Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies. Stroudsburg: ACL, 2021: 39-49.
[23] LI S, JI H, HAN J W. Document-level event argument extraction by conditional generation[C]//Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies. Stroudsburg: ACL, 2021: 894-908.