Aspect Is Not You Need: No-aspect Differential Sentiment Framework for Aspect-based Sentiment Analysis

Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 1599 - 1609 July 10-15, 2022 ©2022 Association for Computational Linguistics Aspect Is Not You Need: No-aspect Differential Sentiment Framework for Aspect-based Sentiment Analysis Jiahao Cao, Rui Liu, Huailiang Peng∗, Lei Jiang, Xu Bai Institute of Information Engineering, Chinese Academy of Sciences School of Cyber Security, University of Chinese Academy of Sciences {caojiahao,liurui3221,penghuailiang,jianglei,baixu}@iie.ac.cn Abstract Aspect-based sentiment analysis (ABSA) is a fine-grained sentiment classification task. Most recent efforts adopt pre-trained model to clas- sify the sentences with aspects. However, the aspect sentiment bias from pre-trained model brings some noise to the ABSA task. Besides, traditional methods using cross-entropy loss are hard to find the potential associations between sentiment polarities. In this work, we analyze the ABSA task from a novel cognition perspec- tive: humans can often judge the sentiment of an aspect even if they do not know what the aspect is. Moreover, it is easier to distinguish positive and negative sentiments than others for human beings because positive and negative are two opposite sentiments. To this end, we pro- pose a no-aspect differential sentiment (NADS) framework for the ABSA task. We first design a no-aspect template by replacing the aspect with a special unbiased character to eliminate the sentiment bias and obtain a stronger rep- resentation. To better get the benefits from the template, we adopt contrastive learning be- tween the no-aspect template and the original sentence. Then we propose a differential sen- timent loss instead of the cross-entropy loss to better classify the sentiments by distinguishing the different distances between sentiments. Our proposed model is a general framework and can be combined with almost all traditional ABSA methods. Experiments on SemEval 2014 show that our framework is still able to predict the sentiment of the aspect even we don’t konw what the aspect is. Moreover, our NADS frame- work boosts three typical ABSA methods and achieves state-of-the-art performance. 1 Introduction Aspect-based sentiment analysis (ABSA) (Jiang et al., 2011) aims to identify the sentiment polarity (i.e., negative, neutral, or positive) of each specific aspect term in a piece of text (Hou et al., 2021; ∗Corresponding author. Dai et al., 2021; Li et al., 2021). For example, in “The food is great, but the service is terrible”, the sentiment towards “food” is positive while the sentiment towards “service” is negative. We need to predict the sentiments of different aspect terms in a sentence. Previous works usually employ pre-trained model to extract the embedding of the concate- nation of the sentence and the aspect term. In this way, the attention mechanism in pre-trained model enhances the connection between aspect and its context (Tang et al., 2016; Song et al., 2019). The experiment results verify its appealing perfor- mance. However, the pre-trained model on large- scale raw corpora tends to internalize aspects’ in- trinsic attributes (Huang et al., 2020) and brings some noise to the ABSA task. For example, for the sentence “Desserts …

Content truncated to 500 words