基于信息融合和数据增强的篇章级事件检测方法

doi:10.3778/j.issn.1673-9418.2312040

摘要/Abstract

摘要： 事件检测是自然语言处理领域的关键任务，旨在识别事件触发词并正确分类其事件类型。语句级事件检测方法未能有效利用文本中的句内和句间事件相关性信息，面临着一词多义、事件共现等众多难题。此外，基于神经网络的事件检测模型需要大量的文本数据作为训练支撑，但语料库的数据不足严重影响着结果的准确率及模型的稳定性。针对上述问题，提出了基于信息融合和数据增强的篇章级事件检测方法LGIA。该方法采用编-解码框架，设计了基于膨胀卷积网络的句子级局部信息抽取模块和基于条件层归一化的篇章级全局信息抽取模块，以深入挖掘整个文档的上下文语义信息和事件间的相关性。同时，采用了同义词替换的数据增强策略，有效扩充了数据样本，从而缓解了数据不足问题带来的影响。经实验验证，LGIA方法在ACE2005数据集上取得了较好的结果，并在数据增强后的TAC-KBP2017数据集上得到了显著的性能提升，F1值分别达到了77.6%和65.3%，相较于现有的基线方法展现出了更优越的性能表现。

关键词: 事件检测, 信息融合, 数据增强, 编码-解码框架

Abstract: Event detection is a key task in the field of natural language processing, aiming to identify event trigger words and correctly classify their event types. Sentence-level event detection methods fail to effectively utilize intra-sentence and inter-sentence event relevance information, facing numerous challenges such as polysemy and event co-occurrence. Additionally, neural network-based event detection models require a large amount of text data for training, but the scarcity of corpus data severely affects the accuracy of results and the stability of the model. To address these issues, this paper proposes a document-level event detection method based on information aggregation and data augmentation, called LGIA (local and global information aggregation). This method adopts an encoder-decoder framework, designing a sentence-level local information extraction module based on dilated convolutional networks and a document-level global information extraction module based on conditional layer normalization, to deeply explore the contextual semantic information and the event correlations of the entire document. Meanwhile, this paper employs a data augmentation strategy of synonym replacement to effectively expand the data samples, thereby alleviating the impact of data scarcity. Experimental results validate that the proposed LGIA method achieves good results on the ACE2005 dataset and significantly improves performance on the augmented TAC-KBP2017 dataset, with F1 scores reaching 77.6% and 65.3%, respectively, demonstrating superior performance compared with existing baseline methods.

Key words: event detection, information aggregation, data augmentation, encoder-decoder framework

谭立君, 胡艳丽, 曹健威, 谭真. 基于信息融合和数据增强的篇章级事件检测方法[J]. 计算机科学与探索, 2024, 18(11): 3015-3026.

TAN Lijun, HU Yanli, CAO Jianwei, TAN Zhen. Document-Level Event Detection Method Based on Information Aggregation and Data Augmentation[J]. Journal of Frontiers of Computer Science and Technology, 2024, 18(11): 3015-3026.

参考文献

[1] 焦磊, 云静, 刘利民, 等. 封闭域深度学习事件抽取方法研究综述[J]. 计算机科学与探索, 2023, 17(3): 533-548.
JIAO L, YUN J, LIU L M, et al. Overview of closed-domain deep learning event extraction methods[J]. Journal of Frontiers of Computer Science and Technology, 2023, 17(3): 533-548.
[2] LOU D F, LIAO Z L, DENG S M, et al. MLBiNet: a cross-sentence collective event detection network[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: 4829-4839.
[3] JI H, GRISHMAN R. Refining event extraction through cross-document inference[C]//Proceedings of the 46th Annual Meeting of the Association for Computational Linguistics, Columbus, Jun 15-20, 2008. Stroudsburg: ACL, 2008: 254-262.
[4] LIAO S S, GRISHMAN R. Using document level cross-event inference to improve event extraction[C]//Proceedings of the 48th Annual Meeting of the Association for Computational Linguistics. Stroudsburg: ACL, 2010: 789-797.
[5] 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.
[6] LIAO J Z, ZHAO X, LI X Y, et al. Learning discriminative neural representations for event detection[C]//Proceedings of the 44th International ACM SIGIR Conference on Research and Development in Information Retrieval. New York: ACM, 2021: 644-653.
[7] CHEN Y B, YANG H, LIU K, et al. Collective event detection via a hierarchical and bias tagging networks with gated multi-level attention mechanisms[C]//Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing. Stroudsburg: ACL, 2018: 1267-1276.
[8] 陈佳丽, 洪宇, 王捷, 等. 利用门控机制融合依存与语义信息的事件检测方法[J]. 中文信息学报, 2020, 34(8): 51-60.
CHEN J L, HONG Y, WANG J, et al. Combination of dependency and semantic information via gated mechanism for event detection[J]. Journal of Chinese Information Processing, 2020, 34(8): 51-60.
[9] DING K, XU L, LIU M, et al. Combing type-aware attention and graph convolutional networks for event detection[J]. Computers, Materials & Continua, 2022, 74(1): 641-654.
[10] JIAO X T, CHEN Z S, CHEN G Y. Event detection model based on graph attention network and graph convolutional network[C]//Proceedings of the 2023 IEEE 6th Information Technology, Networking, Electronic and Automation Control Conference. Piscataway: IEEE, 2023: 322-329.
[11] WU G Q, LU Z Y, ZHUO X R, et al. Semantic fusion enhanced event detection via multi-graph attention network with skip connection[J]. IEEE Transactions on Emerging Topics in Computational Intelligence, 2023, 7(3): 931-941.
[12] LIU X, LUO Z, HUANG H Y. Jointly multiple events extraction via attention-based graph information aggregation[C]//Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing. Stroudsburg: ACL, 2018: 1247-1256.
[13] ZHENG S, CAO W, XU W, et al. Doc2EDAG: an end-to-end document-level framework for Chinese financial event extraction[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: 337-346.
[14] 王雷, 李瑞轩, 李玉华, 等. 文档级无触发词事件抽取联合模型[J]. 计算机科学与探索, 2021, 15(12): 2327-2334.
WANG L, LI R X, LI Y H, et al. Joint model for document-level event extraction without triggers[J]. Journal of Frontiers of Computer Science and Technology, 2021, 15(12): 2327-2334.
[15] VIDAKOVIC J, DOSILOVIC F K, PLUSCEC D. Abstractive summarization as augmentation for document-level event detection[EB/OL]. [2023-10-21]. https://arxiv.org/abs/2305. 18023.
[16] ZHOU J, SHUANG K, AN Z Z, et al. Improving document-level event detection with event relation graph[J]. Information Sciences, 2023, 645: 119355.
[17] YANG S, FENG D W, QIAO L B, et al. Exploring pre-trained language models for event extraction and generation[C]//Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics. Stroudsburg: ACL, 2019: 5284-5294.
[18] WANG X Z, HAN X, LIU Z Y, et al. Adversarial training for weakly supervised event detection[C]//Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies. Stroudsburg: ACL, 2019: 998-1008.
[19] TONG M H, XU B, WANG S, et al. Improving event detection via open-domain trigger knowledge[C]//Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics. Stroudsburg: ACL, 2020: 5887-5897.
[20] DENG S M, ZHANG N Y, LI L Q, et al. OntoED: low-resource event detection with ontology embedding[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: 2828-2839.
[21] SUTSKEVER I, VINYALS O, LE Q V. Sequence to sequence learning with neural networks[C]//Advances in Neural Information Processing Systems 27, Montreal, Dec 8-13, 2014: 3104-3112.
[22] 陆垚杰, 林鸿宇, 韩先培, 等. 基于语言学扰动的事件检测数据增强方法[J]. 中文信息学报, 2019, 33(7): 110-117.
LU Y J, LIN H Y, HAN X P, et al. Linguistic perturbation based data augmentation for event detection[J]. Journal of Chinese Information Processing, 2019, 33(7): 110-117.
[23] 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.
[24] NGUYEN T H, GRISHMAN R. Graph convolutional networks with argument-aware pooling for event detection[C]//Proceedings of the 32nd AAAI Conference on Artificial Intelligence and the 30th Innovative Applications of Artificial Intelligence Conference and the 8th AAAI Symposium on Educational Advances in Artificial Intelligence. Menlo Park: AAAI, 2018: 5900-5907.
[25] LU Y J, LIN H Y, HAN X P, et al. Distilling discrimination and generalization knowledge for event detection via delta-representation learning[C]//Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics.Stroudsburg: ACL, 2019: 4366-4376.
[26] LIU J, CHEN Y F, XU J N. Saliency as evidence: event detection with trigger saliency attribution[C]//Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics. Stroudsburg: ACL, 2022: 4573-4585.
[27] LING T T, CHEN L, SHENG H X, et al. Sentence-level event detection without triggers via prompt learning and machine reading comprehension[EB/OL]. [2023-10-21]. https://arxiv.org/abs/2306.14176.
[28] MHAMDI M, FREEDMAN M, MAY J. Contextualized cross-lingual event trigger extraction with minimal resources[C]//Proceedings of the 23rd Conference on Computational Natural Language Learning, Hong Kong, China, Nov 3-4, 2019. Stroudsburg: ACL, 2019: 656-665.
[29] LI Z Q, HONG Y, HE S M, et al. A mixture of experts with adaptive semantic encoding for event detection[C]//Proceedings of the 2023 International Joint Conference on Neural Networks. Piscataway: IEEE, 2023: 1-7.