Text Emotion Distribution Learning from Small Sample: A Meta-Learning Approach

Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing, pages 3957–3967, Hong Kong, China, November 3–7, 2019. c⃝2019 Association for Computational Linguistics 3957 Text Emotion Distribution Learning from Small Sample: A Meta-Learning Approach Zhenjie Zhao Xiaojuan Ma Department of Computer Science and Engineering Hong Kong University of Science and Technology Clear Water Bay, Hong Kong {zzhaoao, mxj}@cse.ust.hk Abstract Text emotion distribution learning (EDL) aims to develop models that can predict the inten- sity values of a sentence across a set of emo- tion categories. Existing methods based on supervised learning require a large amount of well-labelled training data, which is difficult to obtain due to inconsistent perception of fine-grained emotion intensity. In this paper, we propose a meta-learning approach to learn text emotion distributions from a small sam- ple. Specifically, we propose to learn low-rank sentence embeddings by tensor decomposition to capture their contextual semantic similarity, and use K-nearest neighbors (KNNs) of each sentence in the embedding space to generate sample clusters. We then train a meta-learner that can adapt to new data with only a few training samples on the clusters, and further fit the meta-learner on KNNs of a testing sample for EDL. In this way, we effectively augment the learning ability of a model on the small sample. To demonstrate the performance, we compare the proposed approach with state-of- the-art EDL methods on a widely used EDL dataset: SemEval 2007 Task 14 (Strapparava and Mihalcea, 2007). Results show the superi- ority of our method on small-sample emotion distribution learning. 1 Introduction Analyzing emotions in text automatically is an im- portant topic (Yadollahi et al., 2017) with widely used applications such as classifying e-commerce products reviews (Rao et al., 2016) and devel- oping emotionally intelligent chatbots for health- care (Fadhil and Gabrielli, 2017), to name a few. Text emotion analysis aims to recognize writers’ emotional states towards particular top- ics or subjects (Yadollahi et al., 2017), and has attracted considerable research efforts in the last few decades (Canales and Mart´ınez-Barco, 2014; Abdul-Mageed and Ungar, 2017; Yu et al., 2018; Zhang et al., 2018). Existing works cast text emotion analysis into three types of tasks: single label learning (SLL), multi-label learning (MLL), and label distribution learning (LDL). In SLL, one particular name in an emotion category (referred as label in the fol- lowing) is predicted (Abdul-Mageed and Ungar, 2017), e.g., joyful, angry, or sad. However, due to the correlation among different emotions, one sentence can potentially contain multiple differ- ent emotions (Yu et al., 2018). Therefore, a more practical way is to assign multiple labels to a sen- tence (Yu et al., 2018), namely, MLL. Label dis- tribution learning, which is called emotion distri- bution learning (EDL) in emotion mining (Zhou et al., 2016), goes a step further and assigns one in- tensity value to each emotion, which is necessary to encode fine-grained information for comparing strength of different emotions, etc. In general, there are three …

Content truncated to 500 words