DOER: Dual Cross-Shared RNN for Aspect Term-Polarity Co-Extraction

Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, pages 591–601 Florence, Italy, July 28 - August 2, 2019. c⃝2019 Association for Computational Linguistics 591 DOER: Dual Cross-Shared RNN for Aspect Term-Polarity Co-Extraction Huaishao Luo1, Tianrui Li1∗, Bing Liu2, Junbo Zhang3,4,5 1School of Information Science and Technology, Southwest Jiaotong University, China huaishaoluo@gmail.com, trli@swjtu.edu.cn 2Department of Computer Science, University of Illinois at Chicago, USA liub@uic.edu 3JD Intelligent Cities Business Unit & 4JD Intelligent Cities Research, China 5Institute of Artificial Intelligence, Southwest Jiaotong University, China msjunbozhang@outlook.com Abstract This paper focuses on two related subtasks of aspect-based sentiment analysis, namely as- pect term extraction and aspect sentiment clas- sification, which we call aspect term-polarity co-extraction. The former task is to extract as- pects of a product or service from an opin- ion document, and the latter is to identify the polarity expressed in the document about these extracted aspects. Most existing algo- rithms address them as two separate tasks and solve them one by one, or only perform one task, which can be complicated for real appli- cations. In this paper, we treat these two tasks as two sequence labeling problems and pro- pose a novel Dual crOss-sharEd RNN frame- work (DOER) to generate all aspect term- polarity pairs of the input sentence simultane- ously. Specifically, DOER involves a dual re- current neural network to extract the respective representation of each task, and a cross-shared unit to consider the relationship between them. Experimental results demonstrate that the pro- posed framework outperforms state-of-the-art baselines on three benchmark datasets. 1 Introduction Aspect terms extraction (ATE) and aspect senti- ment classification (ASC) are two fundamental, fine-grained subtasks of aspect-based sentiment analysis. Aspect term extraction is the task of ex- tracting the attributes (or aspects) of an entity upon which opinions have been expressed, and aspect sentiment classification is the task of identifying the polarities expressed on these extracted aspects in the opinion text (Hu and Liu, 2004). Consider the example in Figure 1, which contains com- ments that people expressed about the aspect terms “operating system”, “preloaded software”, “key- board”, “bag”, “price”, and “service” labeled with their polarities, respectively. The polarities contain ∗ Tianrui Li is the corresponding author. I love the [operating system]positiveand the [preloaded software]positive. No backlit [keyboard]conflict, but not an issue for me. You may need to special order a [bag]neutral. The [price]positive is reasonable although the [service]negative is poor. Figure 1: Aspect terms extraction and aspect sentiment classification. four classes, e.g., positive (PO), conflict (CF), neu- tral (NT)1, and negative (NG). To facilitate practical applications, our goal is to solve ATE and ASC simultaneously. For easy description and discussion, these two subtasks are referred to as aspect term-polarity co-extraction. Both ATE and ASC have attracted a great of atten- tion among researchers, but they are rarely solved together at the same time due to some challenges: 1) ATE and ASC are quite different tasks. ATE is an extraction or sequence labeling task (Jakob and Gurevych, 2010; Wang et al., …

Content truncated to 500 words