论文地址：https://aclanthology.org/P19-1096.pdf

背景

情绪原因提取（ECE）旨在提取导致文本中情绪表达的潜在原因。

Figure 1 displays an example from this corpus, There are five clauses in a document.
The emotion "happy" is contained in the fourth clause. We denote this clause as emotion clause, which refers to a clause that contains emotions. It has two corresponding causes: "a policeman visited the old man with the lost money" in the second clause, and "told him that the thief was caught" in the third clause.
We denote them as cause clause, which refers to a clause that contains causes.

Figure 1: An example showing the difference between the ECE task and the ECPE task.

Take Figure 1 for example, given the annotation of emotion: "happy", the goal of ECE is to track the two corresponding cause clauses: "a policeman visited the old man with the lost money" and "and told him that the thief was caught". While in the ECPE task, the goal is to directly extract all pairs of emotion clause and cause clause, including ("The old man was very happy", "a policeman visited the old man with the lost money") and ("The old man was very happy", "and told him that the thief was caught"), without providing the emotion annotation "happy".

问题

1） 在ECE中，在提取原因之前必须对情绪进行注释，这极大地限制了其在现实世界场景中的应用；
2） 先注释情绪，然后提取原因的方法忽略了情绪和原因是相互指示的这一事实。

解决办法

我们提出了一种两步方法来解决这一新的ECPE任务——
步骤1通过两种多任务学习网络将情绪-原因对提取任务转换为两个子任务（分别为情绪提取和原因提取），目的是提取一组情绪从句和一组原因从句。

Figure 2: The Model for Independent Multi-task Learning (Indep).

The lower layer consists of a set of word-level Bi-LSTM modules, each of which corresponds to one clause, and accumulate the context information for each word of the clause. The hidden state of the jth word in the ith clause hi,j is obtained based on a bi-directional LSTM. Attention mechanism is then adopt to get a clause representation si.
The upper layer consists of two components: one for emotion extraction and another for cause extraction. Each component is a clause-level BiLSTM which receives the independent clause representations [s1, s2, ..., s|d|] obtained at the lower layer as inputs. The hidden states of two component Bi-LSTM, re i and rc i , can be viewed as the context-aware representation of clause ci, and finally feed to the softmax layer for emotion prediction and cause predication.

由于提供情绪可以帮助更好地发现原因；了解原因也可能有助于更准确地提取情绪。受此启发，我们进一步提出了an interactive multi-task learning network，作为前一个网络的增强版本，以捕捉情绪和原因之间的相关性。其结构如图3所示。

Figure 3: Two Models for Interactive Multi-task Learning: (a) Inter-EC, which uses emotion extraction to improve cause extraction (b) Inter-CE, which uses cause extraction to enhance emotion extraction.

步骤2执行情感原因配对和过滤。我们通过将笛卡尔乘积应用于情感集E和原因集C，将它们配对。这就产生了一组候选的情绪-原因对。最后训练一个滤波器来消除不包含因果关系的对。

相关工作

Cheng等人（2017）专注于使用多用户结构对中国微博进行原因检测。他们正式化了微博的两个原因检测任务（基于当前子转发的原因检测和基于原始子转发的理由检测），并引入了SVM和LSTM来处理它们。
Chen等人（2018a）提出了一种分层卷积神经网络（Hier-CNN），该网络使用子句级编码器和子转发级编码器分别结合单词上下文特征和基于事件的特征。

实验

数据集
[1]GUI L, WU D, XU R, 等. Event-driven emotion cause extraction with corpus construction[C/OL]//Proceedings of the 2016 Conference on Empirical Methods in Natural Language Processing, Austin, Texas. 2016. http://dx.doi.org/10.18653/v1/d16-1170. DOI:10.18653/v1/d16-1170.

Table 1: The proportion of documents with different number of emotion-cause pairs in the merged dataset.

评价指标
the precision, recall, and F1 score

Table 2: Experimental results of all proposed models and variants using precision, recall, and F1-measure as metrics on the ECPE task as well as the two sub-tasks.

贡献

我们提出了一个新的任务：情绪-原因配对提取（ECPE）。它解决了传统ECE任务在提取原因之前依赖于对情绪的注释的缺点，并允许将情绪原因分析应用于真实世界的场景

我们提出了一个两步框架来解决ECPE任务，该任务首先执行个人情绪提取和原因提取，然后进行情绪-原因配对和过滤。