基于事件表示和对比学习的深度事件聚类方法  

Deep event clustering method based on event representation and contrastive learning

作　　者：蒋小霞 黄瑞章[1,2,3] 白瑞娜 任丽娜 陈艳平 JIANG Xiaoxia;HUANG Ruizhang;BAI Ruina;REN Lina;CHEN Yanping(State Key Laboratory of Public Big Data(Guizhou University),Guiyang Guizhou 550025,China;Text Computing&Cognitive Intelligence Engineering Research Center of National Education Ministry(Guizhou University),Guiyang Guizhou 550025,China;College of Computer Science and Technology,Guizhou University,Guiyang Guizhou 550025,China)

机构地区：[1]公共大数据国家重点实验室(贵州大学),贵阳550025 [2]文本计算与认知智能教育部工程研究中心(贵州大学),贵阳550025 [3]贵州大学计算机科学与技术学院,贵阳550025

出　　处：《计算机应用》2024年第6期1734-1742,共9页journal of Computer Applications

基　　金：国家自然科学基金资助项目(62066007);贵州省教育厅职业教育科研项目(GZZJ-Q2022028);贵州省科技支撑计划项目(黔科合支撑[2023]一般300)。

摘　　要：针对现有深度聚类方法不考虑事件信息及其结构特点而难以有效划分事件类型的问题,提出一种基于事件表示和对比学习的深度事件聚类方法(DEC_ERCL)。首先,利用信息识别手段从非结构化文本中识别结构化的事件信息,避免冗余信息对事件语义的影响;其次,将事件的结构信息集成于自编码器中学习低维稠密的事件表示,并以此作为下游聚类划分的依据;最后,为有效建模事件之间的细微差异,在特征学习过程中加入多正例对比损失。在数据集DuEE、FewFC、Military和ACE2005上的实验结果表明,相较于其他深度聚类方法,所提方法在准确率和标准化互信息(NMI)评价指标上均表现更好;相较于次优的方法,DEC_ERCL的聚类准确率分别提升了17.85%、9.26%、7.36%和33.54%,表明了DEC_ERCL具有更好的事件聚类效果。Aiming at the problem that the existing deep clustering methods can not efficiently divide event types without considering event information and its structural characteristics,a Deep Event Clustering method based on Event Representation and Contrastive Learning(DEC_ERCL)was proposed.Firstly,information recognition was utilized to identify structured event information from unstructured text,thus the impact of redundant information on event semantics was avoided.Secondly,the structural information of the event was integrated into the autoencoder to learn the low-dimensional dense event representation,which was used as the basis for downstream clustering.Finally,in order to effectively model the subtle differences between events,a contrast loss with multiple positive examples was added to the feature learning process.Experimental results on the datasets DuEE,FewFC,Military and ACE2005 show that the proposed method performs better than other deep clustering methods in accuracy and Normalized Mutual Information(NMI)evaluation indexes.Compared with the suboptimal method,the accuracy of DEC_ERCL is increased by 17.85%,9.26%,7.36%and 33.54%,respectively,indicating that DEC_ERCL has better event clustering effect.

关 键 词：深度聚类 文本聚类 事件表示 事件结构 对比学习 

分 类 号：TP391.1[自动化与计算机技术—计算机应用技术]