A Knowledge-Integrate Cross-Domain Data Generation Method for Aspect and Opinion Co-Extraction  

作　　者：Hao Zhang Yegang Li Jiachen Yang Rujiang Bai Hao Zhang;Yegang Li;Jiachen Yang;Rujiang Bai(School of Computer Science and Technology, Shandong University of Technology, Zibo, China;Institute of Information Management, Shandong University of Technology, Zibo, China)

机构地区：[1]School of Computer Science and Technology, Shandong University of Technology, Zibo, China [2]Institute of Information Management, Shandong University of Technology, Zibo, China

出　　处：《Journal of Computer and Communications》2023年第12期31-48,共18页电脑和通信（英文）

摘　　要：To address the difficulty of training high-quality models in some specific domains due to the lack of fine-grained annotation resources, we propose in this paper a knowledge-integrated cross-domain data generation method for unsupervised domain adaptation tasks. Specifically, we extract domain features, lexical and syntactic knowledge from source-domain and target-domain data, and use a masking model with an extended masking strategy and a re-masking strategy to obtain domain-specific data that remove domain-specific features. Finally, we improve the sequence generation model BART and use it to generate high-quality target domain data for the task of aspect and opinion co-extraction from the target domain. Experiments were performed on three conventional English datasets from different domains, and our method generates more accurate and diverse target domain data with the best results compared to previous methods.To address the difficulty of training high-quality models in some specific domains due to the lack of fine-grained annotation resources, we propose in this paper a knowledge-integrated cross-domain data generation method for unsupervised domain adaptation tasks. Specifically, we extract domain features, lexical and syntactic knowledge from source-domain and target-domain data, and use a masking model with an extended masking strategy and a re-masking strategy to obtain domain-specific data that remove domain-specific features. Finally, we improve the sequence generation model BART and use it to generate high-quality target domain data for the task of aspect and opinion co-extraction from the target domain. Experiments were performed on three conventional English datasets from different domains, and our method generates more accurate and diverse target domain data with the best results compared to previous methods.

关 键 词：Knowledge-Integrate Domain Adaptation Text Generation Aspect and Opinion Co-Extraction 

分 类 号：O17[理学—数学]