Unified Feature and Instance Based Domain Adaptation for Aspect-Based Sentiment Analysis

Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing, pages 7035–7045, November 16–20, 2020. c⃝2020 Association for Computational Linguistics 7035 Unified Feature and Instance Based Domain Adaptation for Aspect-Based Sentiment Analysis Chenggong Gong∗, Jianfei Yu∗, and Rui Xia† School of Computer Science and Engineering, Nanjing University of Science and Technology, China {cggong, jfyu, rxia}@njust.edu.cn Abstract The supervised models for aspect-based sen- timent analysis (ABSA) rely heavily on la- beled data. However, fine-grained labeled data are scarce for the ABSA task. To alleviate the dependence on labeled data, prior works mainly focused on feature-based adaptation, which used the domain-shared knowledge to construct auxiliary tasks or domain adversarial learning to bridge the gap between domains, while ignored the attribute of instance-based adaptation. To resolve this limitation, we pro- pose an end-to-end framework to jointly per- form feature and instance based adaptation for the ABSA task in this paper. Based on BERT, we learn domain-invariant feature representa- tions by using part-of-speech features and syn- tactic dependency relations to construct aux- iliary tasks, and jointly perform word-level instance weighting in the framework of se- quence labeling. Experiment results on four benchmarks show that the proposed method can achieve significant improvements in com- parison with the state-of-the-arts in both tasks of cross-domain End2End ABSA and cross- domain aspect extraction. 1 Introduction Aspect extraction and aspect sentiment classifica- tion are two important sub-tasks in Aspect Based Sentiment Analysis (ABSA) (Liu, 2012; Pontiki et al., 2016), which aim to extract aspect terms and predict the sentiment polarities of the given aspect terms, respectively. Since these two sub-tasks have been well studied in the literature, a number of re- cent studies focus on the End2End ABSA task by employing a unified tagging scheme to tackle the two sub-tasks in an end-to-end manner (Mitchell et al., 2013; Zhang et al., 2015; Li et al., 2019a). The unified tagging scheme fuses aspect boundary ∗Equal contribution. † Corresponding author. tags {B, I, O} and sentiment polarities {POS, NEG, NEU} together, and formulates End2End ABSA as a sequence labeling problem. For example, given a sentence “The price is reasonable, although the service is poor.”, the End2End ABSA task aims to jointly extract aspect terms and detect sentiment polarities over them. The extracted pairs in this ex- ample are {“price”: Positive; “service”: Negative}. However, these existing studies heavily rely on su- pervised learning over a large amount of labeled data, which is usually hard to obtain for ABSA due to the intensive nature of human annotation. Therefore, it will be very attractive to explore the End2End ABSA task in a cross-domain setting, which allows us to train a robust ABSA model for a resource-poor target domain based on enough annotated data in a resource-rich source domain. Traditional domain adaptation methods primar- ily focus on coarse-grained sentiment classifica- tion (Blitzer et al., 2007; Pan et al., 2010; Glorot et al., 2011; Bollegala et al., 2012; Xia et al., 2013; Yu and Jiang, 2016; Ganin et al., 2016; Li et al., 2018b). …

Content truncated to 500 words