Abstract：Most existing deep learning emotion cause extraction methods are unable to model latent semantic relationships between clauses. In addition, these methods are not easily controlled, are difficult to interpret and need high-quality annotations. This paper presents an emotion cause extraction method that incorporates rule distillation with a hierarchical attention network. The hierarchical attention network uses position encoding and the residual structure to capture the latent semantic relationships within the clauses and between the clauses and the emotional expression. A knowledge distillation architecture based on adversarial learning then introduces linguistic rules related to the emotion cause expression into the deep neural network. Tests on a Chinese emotion cause extraction dataset show that this method outperforms the state-of-the-art method by 0.02 in F1, the best known result.
[1] LEE S Y M, CHEN Y, HUANG C R. A text-driven rule-based system for emotion cause detection[C]//Proceedings of the NAACL HLT 2010 Workshop on Computational Approaches to Analysis and Generation of Emotion in Text. Los Angeles, USA:Association for Computational Linguistics, 2010:45-53. [2] LEE S Y M, CHEN Y, HUANG C R, et al. Detecting emotion causes with a linguistic rule-based approach[J]. Computational Intelligence, 2013, 29(3):390-416. [3] GUI L, YUAN L, XU R F, et al. Emotion cause detection with linguistic construction in Chinese Weibo Text[C]//Proceedings of the Natural Language Processing and Chinese Computing. Shenzhen, China:Springer, 2014:457-464. [4] GUI L, HU J N, HE Y L, et al. A question answering approach to emotion cause extraction[C]//Proceedings of the Conference on Empirical Methods in Natural Language Processing. Copenhagen, Denmark:Association for Computational Linguistics, 2017:1593-1602. [5] DING Z X, HE H H, ZHANG M R, et al. From independent prediction to re-ordered prediction:Integrating relative position and global label information to emotion cause identification[C]//Proceedings of the National Conference on Artificial Intelligence. Hawaii, USA:AAAI, 2019:6343-6350. [6] GUI L, WU D Y, XU R F, et al. Event-driven emotion cause extraction with corpus constructions[C]//Proceedings of 2016 Conference on Empirical Methods in Natural Language Processing. Austin, USA:Association for Computational Linguistics, 2016:1639-1649. [7] 余传明, 李浩男, 安璐. 基于多任务深度学习的文本情感原因分析[J]. 广西师范大学学报(自然科学版), 2019, 37(1):50-61.YU C M, LI H N, AN L. Analysis of text emotion cause based on multi-task deep learning[J]. Journal of Guangxi Normal University (Natural Science Edition), 2019, 37(1):50-61. (in Chinese) [8] HINTON G, VINYALS O, DEAN J. Distilling the knowledge in a neural network[C]//Proceedings of the Workshops of Advances in Neural Information Processing Systems. Montreal, Canada, 2014:200-208. [9] HU Z T, MA X Z, LIU Z Z, et al. Harnessing deep neural networks with logic rules[C]//Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics. Berlin, Germany:Association for Computational Linguistics, 2016:2410-2420. [10] LIU J, CHEN Y B, LIU K. Exploiting the ground-truth:An adversarial imitation based knowledge distillation approach forevent detection[C]//Proceedings of the Thirty-Third AAAI Conference on Artificial Intelligence. Hawaii, USA:AAAI, 2019:6754-6761. [11] KRISHNA K, JYOTHI P, IYYER M. Revisiting theimportance of encoding logic rules in sentiment classification[C]//Proceedings of 2018 Conference on Empirical Methods in Natural Language Processing. Brussels, Belgium:Association for Computational Linguistics, 2018:4743-4751. [12] RUSSO I, CASELLI T, RUBINO F, et al. EMOCause:An easy-adaptable approach to emotion cause contexts[C]//Proceedings of the 2nd Workshop on Cmputational Approaches to Subjectivity and Sentiment Analysis. Portland, USA:Association for Computational Linguistics, 2011:153-160. [13] LI X J, SONG K S, FENG S, et al. A co-attention neural network model for emotion cause analysis with emotional context awareness[C]//Proceedings of 2018 Conference on Empirical Methods in Natural Language Processing. Brussels, Belgium:Association for Computational Linguistics, 2018:4752-4757. [14] YANG Z C, YANG D Y, DYER C, et al. Hierarchical attention networks for document classification[C]//Proceedings of 2016 Conference of the North American Chapter of the Association for Computational Linguistics. San Diego, USA:Association for Computational Linguistics, 2016:1480-1489.
2020, Vol. 60 
