AX-MABSA: A Framework for Extremely Weakly Supervised Multi-label Aspect Based Sentiment Analysis

Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing, pages 6136 - 6147 December 7-11, 2022 ©2022 Association for Computational Linguistics AX-MABSA: A Framework for Extremely Weakly Supervised Multi-label Aspect Based Sentiment Analysis Sabyasachi Kamila1, Walid Magdy1, Sourav Dutta2 and MingXue Wang2 1School of Informatics, University of Edinburgh 2Huawei Research Centre, Dublin, Ireland {skamila,wmagdy}@inf.ed.ac.uk, {sourav.dutta2,wangmingxue1}@huawei.com Abstract Aspect Based Sentiment Analysis is a domi- nant research area with potential applications in social media analytics, business, finance, and health. Prior works in this area are primarily based on supervised methods, with a few tech- niques using weak supervision limited to pre- dicting a single aspect category per review sen- tence. In this paper, we present an extremely weakly supervised multi-label Aspect Category Sentiment Analysis framework which does not use any labelled data. We only rely on a sin- gle word per class as an initial indicative in- formation. We further propose an automatic word selection technique to choose these seed categories and sentiment words. We explore unsupervised language model post-training to improve the overall performance, and propose a multi-label generator model to generate mul- tiple aspect category-sentiment pairs per re- view sentence. Experiments conducted on four benchmark datasets showcase our method to outperform other weakly supervised baselines by a significant margin.1 1 Introduction Aspect-based sentiment analysis (ABSA) is a well- known sentiment analysis task which provides more fine-grained information than simple senti- ment understanding (Liu, 2012). The main goal of ABSA is to find the aspects and its associated sentiment within a given text. While the works on ABSA have expanded in different directions, it has primarily two sub-tasks, Aspect Term Sentiment Analysis (ATSA) and Aspect Category Sentiment Analysis (ACSA) (Xue and Li, 2018). ATSA con- sists of different tasks like aspect term extraction (Li et al., 2018; Luo et al., 2019; Li et al., 2020a; Shi et al., 2021), aspect term sentiment classifica- tion (He et al., 2018; Chen and Qian, 2019; Hou et al., 2021), opinion term extraction (Dai and Song, 1Code, data and model are available at https://github. com/sabyasachi-kamila/AX-MABSA 2019; He et al., 2019; Chen and Qian, 2020b), aspect-oriented opinion term extraction (Fan et al., 2019; Wu et al., 2020a), aspect-opinion pair extrac- tion (Zhao et al., 2020), etc. For example, in the sentence “The sushi is top-notch, the waiter is at- tentive, but the atmosphere is dull.", ATSA would extract the aspect terms ‘sushi’, ‘waiter’ and ‘atmo- sphere’; opinion terms ‘top-notch’, ‘attentive’, and ‘dull’; and their associated sentiments ‘positive’, ‘positive’ and ‘negative’. The other sub-task ACSA aims to find the higher order aspect categories and its associated sentiment from a given text. In the above example, ACSA would detect the categories as ‘food’ (as ‘pasta’ is a type of ‘food’), ‘service’ and ‘ambience’; and the associated sentiments as ‘positive’, ‘positive’ and ‘negative’. Existing research on ABSA is dominated by su- pervised methods, where labeled training data is provided (Chen et al., 2017; Xue and Li, 2018; Cai et al., 2021; Liu et al., 2021; Xu …

Content truncated to 500 words