Interpretable Relevant Emotion Ranking with Event-Driven Attention

Abstract

Multiple emotions with different intensities are often evoked by events described in documents. Oftentimes, such event information is hidden and needs to be discovered from texts. Unveiling the hidden event information can help to understand how the emotions are evoked and provide explainable results. However, existing studies often ignore the latent event information. In this paper, we proposed a novel interpretable relevant emotion ranking model with the event information incorporated into a deep learning architecture using the event-driven attentions. Moreover, corpuslevel event embeddings and document-level event distributions are introduced respectively to consider the global events in corpus and the document-specific events simultaneously. Experimental results on three real-world corpora show that the proposed approach performs remarkably better than the state-of-the-art emotion detection approaches and multi-label approaches. Moreover, interpretable results can be obtained to shed light on the events which trigger certain emotions.

Hashtags, Emotions, and Comments: A Large-Scale Dataset to Understand Fine-Grained Social Emotions to Online Topics

“…Responsive emotions are collected from limited readers and hence cannot reflect social emotions from the public. Other studies adopt emotion votes on news websites to gather public feelings on news (Li et al, 2016;Zhou et al, 2018;Yang et al, 2019). However, news reports usually exhibit more formal style than online topics in social media language.…”

“…Nevertheless, most of the related work focus on the feelings from writers (Tang et al, 2014;Huang and Carley, 2019;Singh et al, 2019) and the existing studies concerning reader emotions mostly tackle well-written texts, such as news reports (Li et al, 2016;Zhou et al, 2018;Yang et al, 2019). Limited work has been done to characterize collective feelings from the public (henceforth social emotions) to an online topic described with fragmented and colloquial social media language.…”

“…Its goal is to predict the possible emotions from the majority of readers given a few words description of a trending social media topic. The task will benefit various applications, such as topic analysis (Wang et al, 2011) and event detection (Yang et al, 2019). It would help people understand and foresee how the general public thinks of an event even before it becomes a trending topic on social media.…”

Embedding Refinement Framework for Targeted Aspect-Based Sentiment Analysis

“…Recently, the attention mechanism has achieved a great success in many NLP tasks [34], [35], [36], [37], [38], [39], [40], [41], [42]. In the task of (T)ABSA, attention mechanism can help models effectively distinguishing the sentiment polarities of different aspects in the same sentence [9], [10], [11], [43], [44], [45], [46], [47], [48], [49].…”

MAVEN: A Massive General Domain Event Detection Dataset

“…For instance, in the sentence "Bill Gates founded Microsoft in 1975", an ED model should recognize that the word "founded" is the trigger of a Found event. ED is the first stage to extract event knowledge from text (Ahn, 2006) and also fundamental to various NLP applications (Yang et al, 2003;Basile et al, 2014;Cheng and Erk, 2018;Yang et al, 2019).…”