Conditional Augmentation for Aspect Term Extraction via Masked Sequence-to-Sequence Generation

Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 7056–7066 July 5 - 10, 2020. c⃝2020 Association for Computational Linguistics 7056 Conditional Augmentation for Aspect Term Extraction via Masked Sequence-to-Sequence Generation Kun Li1, Chengbo Chen1, Xiaojun Quan1∗, Qing Ling1, Yan Song2 1School of Data and Computer Science, Sun Yat-sen University, China 2Sinovation Ventures {likun36, chenchb7}@mail2.sysu.edu.cn {quanxj3, lingqing556}@mail.sysu.edu.cn, clksong@gmail.com Abstract Aspect term extraction aims to extract aspect terms from review texts as opinion targets for sentiment analysis. One of the big challenges with this task is the lack of sufficient annotated data. While data augmentation is potentially an effective technique to address the above issue, it is uncontrollable as it may change aspect words and aspect labels unexpectedly. In this paper, we formulate the data augmen- tation as a conditional generation task: gen- erating a new sentence while preserving the original opinion targets and labels. We pro- pose a masked sequence-to-sequence method for conditional augmentation of aspect term extraction. Unlike existing augmentation ap- proaches, ours is controllable and allows us to generate more diversified sentences. Exper- imental results confirm that our method allevi- ates the data scarcity problem significantly. It also effectively boosts the performances of sev- eral current models for aspect term extraction. 1 Introduction Aspect term extraction (ATE), which aims to iden- tify and extract the aspects on which users express their sentiments (Hu and Liu, 2004; Liu, 2012), is a fundamental task in aspect-level sentiment analysis. For example, in the sentence of “The screen is very large and crystal clear with amazing colors and resolution”, “screen”, “colors” and “resolution” are the aspect terms to extract in this task. ATE is typically formulated as a sequence la- beling problem (Xu et al., 2018, 2019; Li et al., 2018), where each word is appended with a label indicating if it identifies an aspect. Sentence and label sequence are both used to train a ATE model. One of the remaining challenges with this task is the shortage of annotated data. While data augmen- tation appears to be a solution to this problem, it ∗Corresponding author. ������������: I ������������: The ������������: I �match the original label sequence, and �screen� is the opinion target in this sentence �fail to match the original label sequence ��screen� is not an opinion target in this sentence ���������������: ������������������: O O OIt O O B screen . poor a is is poor so . screen screen . this like really a screen . bought have Figure 1: Examples of ATE augmentation, where B, I and O denote that a word is the beginning, inside and outside of opinion target, respectively. faces two main obstacles here. First, the new sen- tences must adhere to their original label sequences strictly. As shown in Figure 1, the generation A is an effective augmentation as the original label se- quence is preserved, whereas B is not even though it can be a valid review. Second, a noun phrase is regarded as aspect term only if it is …

Content truncated to 500 words