[Announced by Harbin Institute of Technology, Shenzhen] (School of Computer Science and Technology Text/Photo) Recently, the master's thesis Adversarial Training based Cross-lingual Emotion Cause Extraction from the School of Computer Science and Technology, Harbin Institute of Technology, Shenzhen was awarded the Best Paper Award of the 20th International Conference on Computational Linguistics and Intelligent Text Processing, CICLING 2019, of which Harbin Institute of Technology, Shenzhen was the first completion unit and correspondent unit, Yan Hongyu, a 2017 master's degree student from the School of Computer Science and Technology, was the first author, and Professor Xu Ruifeng and Du Jiachen, a 2015 doctoral student, were co-correspondent authors.

This paper focuses on how to automatically identify the factors or events that stimulate emotion in the text with emotional expression. In view of the disadvantages of existing methods, such as high dependence on single language tagging corpus and too many limitations of traditional cross-language methods, this paper proposes a cross-language emotional cause discovery method based on confrontation training, which comprehensively uses different language tagging corpus to improve the performance of emotional cause discovery in the target language. (Editor Wu Ruichan)