【论文阅读】A Comprehensive Survey on Schema-based Event Extraction with Deep Learning

【论文阅读】A Comprehensive Survey on Schema-based Event Extraction with Deep Learning

Qian Li et al. 2021

总结：

• 分类整理了deep learning事件抽取的进展和方法

1. Event Extraction terminologies 定义
   1. Entity 实体: The entity is an object or group of objects in a semantic category. Entity mainly includes people, organizations, places, times, things, etc.
   2. Event mentions 事件提及: The phrase or sentences that describe the event contains a trigger and corresponding arguments.
   3. Event type 事件类型: The event type describes the nature of the event and refers to the category to which the event corresponds, usually represented by the type of the event trigger.
   4. Event trigger 事件触发词: Event trigger refers to the core unit in event extraction, a verb or a noun. Trigger identification is a key step in pipeline-based event extraction.
   5. Event argument 事件论元: Event argument is the main attribute of events. It includes entities, nonentity participants, and time, and so on.
   6. Argument role 论元角色: An argument role is a role played by an argument in an event, that is, the relationship representation between the event arguments and the event triggers.
2. Schema-based event extraction includes four sub-tasks 四个子任务
   1. Event classification 事件分类: Event classification is to determine whether each sentence is an event. Furthermore, if the sentence is an event, we need to determine one or several events types the sentence belongs to. Therefore, the event classification subtask can be seen as a multi-label text classification task.
   2. Trigger identification 识别触发词: It is generally considered that the trigger is the core unit in event extraction that can clearly express an event’s occurrence. The trigger identification subtask it to find the trigger from the text.
   3. Argument identification 识别论元: Argument identification is to identify all the arguments contained in an event type from the text. Argument identification usually depends on the result of event classification and trigger identification.
   4. Argument role classification 论元角色分类: Argument role classification is based on the arguments contained in the event extraction schema, and the category of each argument is classified according to the identified arguments. Thus, it also can be seen as a multi-label text classification task.
3. Event Extraction Paradigm
   1. Pipiline-based 管道模型 代表模型 DMCNN 2015
      1. 把所有子任务视为独立的分类任务treats all sub-tasks as independent classification problems
      2. 将整个EE任务转化为多阶段的分类任务converts event extraction tasks into a multi-stage classification problem
      3. 需要的分类器包括
         1. Trigger classifier 触发词分类器
         2. Argument classifier论元分类器
         3. Argument role classifier 论元角色分类器
      4. +：通过之前的子任务为后续的子任务提供额外的信息，并利用子任务之间的依赖关系 provide additional information for subsequent sub-tasks through previous sub-tasks, and take advantage of dependencies between subtasks
      5. –：错误传递传播error propagation触发词识别错误结果会误导后续任务The trigger can provide additional information for event classification, but the result of wrong trigger identification can also affect event classification
      6. –：缺少子任务之间的interaction和impact。 because the trigger detection always precedes the argument detection, the argument won’t be considered while detecting triggers. Therefore, each link is independent and lacks interaction, ignoring the impact between them.
   2. Joint-based paradigm 联合模型

联合方法构建联合学习模型来触发识别和参数识别，触发器和参数可以相互促进彼此的提取效果。In order to overcome the shortcomings of the pipeline method, researchers proposed a joint method. The joint method constructs a joint learning model to trigger recognition and argument recognition, where the trigger and argument can mutually promote each other’s extraction effect.

1. 事件抽取模型 EE Models   
   1. Traditional EE Models
      1. Pattern matching methods, mainly based on syntax trees or regular expressions
      2. Based on statistics (ML)
   2. 1. Deep Leaning based models
         1. CNN-based
         2. RNN-based
         3. Attention-based
         4. GCN-based
         5. Transformer-based 目前主流 由bert带来的预训练模型（句子级别）
            1. PLMEE 解决角色overlap问题，将不同论元角色分开预测；两阶段都基于bert：trigger和argument
            2. DYGIE++ 多句bert编码 句子之间的上下文
            3. QA方法 句子级
            4. GAIL 使用GAN 强化学习
            5. Transfer learning
         6. 总结：

传统的事件提取方法大多采用人工构建的方法进行特征表示，并使用分类模型对触发器进行分类并识别参数的作用。 近年来，深度学习在图像处理、语音识别、自然语言处理等方面表现出突出的效果。针对传统方法的弊端，系统探讨了基于深度学习的事件提取。 在BERT模型出现之前，主流的方法是从文本中寻找触发器，根据触发器判断文本的事件类型。 最近，随着BERT引入事件抽取模型，基于全文识别事件类型的方法成为主流。 这是因为BERT具有突出的上下文表示能力，在文本分类任务中表现良好，尤其是在数据量很小的情况下。

Most of the traditional event extraction methods adopt the artificial construction method for feature representation and use the classification model to classify triggers and identify the role of the argument. In recent years, the deep learning has shown outstanding effects in image processing, speech recognition, and natural language processing, etc. To settle the drawbacks of traditional methods, deep learning-based event extraction is systematically discussed. Before the emergence of the BERT model, the mainstream method is to find the trigger from the text and judge the event type of the text according to the trigger. Recently, with the introduction of the event extraction model by BERT, the method of identifying event types based on the full text has become mainstream. It is because BERT has outstanding contextual representation ability and performs well in text classification tasks, especially when there is only a small amount of data.

1. 评价指标metrics和量化的结果
   1. Precision 判断为该类中实际为该类的比例
   2. Recall实际为该类中判断为该类的比例
   3. F1 score 是Precision和Recall加权调和平均2*P*R/(P+R)
2. 未来研究方向
   1. Bert预训练的语言问题不同于ee中想要extract的信息，making the dependency relationship between the event arguments?
   2. 缺少标注数据集 人工标注太费事费力 the construction of large scale event extraction data set or the design of automatic construction EE dataset
   3. 联合表示 多任务
   4. 利用外部资源 当前dataset太小的痛点
   5. 篇章级EE 多事件
   6. 开放式EE
3. 总结

本文主要介绍了现有的用于事件提取任务的深度学习模型。与传统方法相比，结论如下：1）基于深度学习的事件提取方法可以自主学习，可以自主学习特征。触发器分类和参数角色分类的性能优于传统方法。 2）随着深度学习的快速发展，机器学习和基于神经网络的深度学习在事件提取方面不断取得良好进展。使用深度学习模型解决缺失数据将为后续研究提供重要的研究方向。首先从事件抽取的三个方面介绍概念和定义。然后我们将基于深度学习的事件提取范式分为管道和关节部分，分别进行介绍。基于深度学习的模型通过改进表示学习方法、模型结构以及附加数据和知识来提高性能。然后，我们介绍带有汇总表和评估指标的数据集。此外，我们在 ACE 2005 数据集的汇总表中给出了领先模型的定量结果。最后，我们总结了事件提取未来可能的研究趋势。This paper principally introduces the existing deep learning models for event extraction tasks. Comparing with the traditional methods, the conclusions are as follows: 1) The event extraction methods based on deep learning can autonomously learn, and they can autonomously learn the features. The performance of trigger classification and the argument role classification is better than the traditional methods. 2) With the rapid development of deep learning, machine learning and deep learning based on neural networks are making good progress in event extraction continuously. Using deep learning models to solve missing data will provide an essential research direction for the followup research. Firstly, we introduce concepts and definitions from three aspects of event extraction. Then we divide the deep learning-based event extraction paradigm into the pipeline and joint parts and introduce them, respectively. Deep learning-based models enhance performance by improving the presentation learning method, model structure, and additional data and knowledge. Then, we introduce the datasets with a summary table and evaluation metrics. Furthermore, we give the quantitative results of the leading models in a summary table on ACE 2005 datasets. Finally, we summarize the possible future research trends of event extraction.