Superbizarre Is Not Superb: Derivational Morphology Improves BERT’s Interpretation of Complex Words

Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing, pages 317–327 August 1–6, 2021. ©2021 Association for Computational Linguistics 317 Bridge-Based Active Domain Adaptation for Aspect Term Extraction Zhuang Chen, Tieyun Qian∗ School of Computer Science, Wuhan University, China {zhchen18, qty}@whu.edu.cn Abstract As a fine-grained task, the annotation cost of aspect term extraction is extremely high. Re- cent attempts alleviate this issue using domain adaptation that transfers common knowledge across domains. Since most aspect terms are domain-specific, they cannot be transferred di- rectly. Existing methods solve this problem by associating aspect terms with pivot words (we call this passive domain adaptation be- cause the transfer of aspect terms relies on the links to pivots). However, all these methods need either manually labeled pivot words or expensive computing resources to build asso- ciations. In this paper, we propose a novel active domain adaptation method. Our goal is to transfer aspect terms by actively supple- menting transferable knowledge. To this end, we construct syntactic bridges by recognizing syntactic roles as pivots instead of as links to pivots. We also build semantic bridges by re- trieving transferable semantic prototypes. Ex- tensive experiments show that our method sig- nificantly outperforms previous approaches. 1 Introduction Aspect term extraction (ATE) is a fundamental task in aspect-based sentiment analysis. Given a re- view sentence “The pizza here is also absolutely delicious.”, ATE aims to extract the term pizza. Re- cent studies define ATE as a sequence tagging task and propose supervised taggers (Wang et al., 2017; Xu et al., 2018). However, due to the high cost of token-level annotation, the lack of labeled data becomes the main obstacle (Chen and Qian, 2019). To alleviate the data deficiency issue, unsuper- vised domain adaptation is proposed to transfer knowledge from the labeled source domain to the unlabeled target domain. Since ATE is a token- level task, it is natural to conduct token-level do- main adaptation. Then a problem arises: many *Corresponding author. ���� ���� ���� ���� ����� ����� � �� �� ��� ��� ��� ��� ��� ��� ���������� ������������� �������������������������������������� Figure 1: The proportion of source aspect terms that appear in target data. R (Restaurant), L (Laptop), and D (Device) are three datasets from different domains. aspect terms are domain-specific and cannot be transferred directly. We present the proportion of source aspect terms that also appear in target test data in Figure 1. As can be seen, in distant trans- fer pairs like R→L, only less than 10% of source aspect terms have appeared in target data. Even in a close pair L→D, the proportion is no more than 40%. In other words, there is a wide discrep- ancy between the data from different domains, and many aspect terms have to be transferred under the guidance of proper references. To solve this problem, previous studies try to associate aspect terms with specific pivot words1. We name these methods passive domain adaptation because the transfer of aspect …

Content truncated to 500 words