Adversarial Attention Modeling for Multi-dimensional Emotion Regression

Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, pages 471–480 Florence, Italy, July 28 - August 2, 2019. c⃝2019 Association for Computational Linguistics 471 Adversarial Attention Modeling for Multi-dimensional Emotion Regression Suyang Zhu, Shoushan Li, Guodong Zhou∗ Natural Language Processing Lab, School of Computer Science and Technology, Soochow University, China syzhu@stu.suda.edu.cn, {lishoushan, gdzhou}@suda.edu.cn Abstract In this paper, we propose an Adversarial Attention Network for the task of multi- dimensional emotion regression, which auto- matically rates multiple emotion dimension scores for an input text. Especially, to deter- mine which words are valuable for a particular emotion dimension, an attention layer is learnt to weight the words in an input sequence. Furthermore, adversarial training is employed between two attention layers to learn better word weights via a discriminator. In particu- lar, a shared attention layer is incorporated to learn public word weights between two emo- tion dimensions. Empirical evaluation on the EMOBANK corpus shows that our approach achieves notable improvements in r-values on both EMOBANK Reader’s and Writer’s multi-dimensional emotion regression tasks in all domains over the state-of-the-art baselines. 1 Introduction Emotion analysis aims to recognize human emo- tion expression in a given text (Mishne et al., 2005; Abdul-Mageed and Ungar, 2017). Typically, stud- ies in emotion analysis can be divided into either emotion classification (Yang et al., 2007; Tripathi et al., 2017) or emotion regression (Yu et al., 2015; Wang et al., 2016a). While emotion classification aims to label an input text with a single or multi- ple emotion categories, emotion regression aims to rate a single or multiple emotion dimension scores of an input text through machine learning models. In this study, we focus on emotion regression. Compared with enormous studies in emotion classification, studies in emotion regression have a late start much due to the inherent difficulty of the regression task and the lack of large-scale emotion regression corpora in high quality. De- spite of its difficulty, emotion regression is more ∗Corresponding author Sample Text: I was very scared when the gunner started shooting the crowd. What a dis- aster! Emotion dimension scores: Valence = 2.0, Arousal = 4.4, Dominance = 2.1 Figure 1: An example of multi-dimensional emotion regression. The dimensional emotion score ranges from 1.0 to 5.0. In this example, the word very in blue only suggests one emotion dimension (i.e, a high Arousal score). The word scared and disaster in red suggest two emotion dimensions. Specifically, scared suggests a low Valence score and a low Dominance score, while Disaster denotes a low Valence score and a high Arousal score. suitable for fine-grained emotion analysis and has gained an increasing attention recently due to the availability of several emotion regression corpora in the last few years (Preotiuc-Pietro et al., 2016; Yu et al., 2016; Hahn and Buechel, 2017). In principle, these emotion regression corpora ap- ply the widely-admitted Valence-Arousal model or Valence-Arousal-Dominance model (Barrett, 2006) to describe emotions with a continuous real number space in two or three dimensions. More- over, while different …

Content truncated to 500 words