近年来,情绪分析方法的研究得到了飞跃式的进展,但作为情绪分析研究任务之一的情绪回归任务因语料的匮乏,目前还没有取得突破性的成果。相比情绪分类的研究,情绪回归方法受分类体系的影响较小,更具有泛化性。该文提出了一种基于维度-标签信息的多元情绪回归方法,可以同时预测输入文本在极性、强度和可控性三个维度的分值。该方法利用情绪维度和情绪类别的互信息,具体的方法是尽可能最大化两个不同情绪标签的文本在表示空间中的距离,从而输出与真实值更接近的预测分数。在英文数据集EMOBANK上的实验结果表明,该方法在均方误差和皮尔森相关系数两个指标上取得了显著提升,尤其是在极性和强度这两个维度上有较好的性能表现。
Abstract
In recent years, emotion analysis has experienced rapid development. As one of the tasks of emotion analysis, the emotion regression is more generalized and less affected by the classification taxonomy, lacking of sufficient corpus, though. In this paper, we propose a multi-dimensional emotion regression method via dimension-label information to predict the input text scores in three dimensions (Valence, Arousal, Dominance). This method conducts emotion regression by the probability of emotion classification prediction, with an objective to maximize the distance between two texts with different emotion labels. Experimental results on EMOBANK show that the proposed method has achieved significant improvement according to the mean square error and Pearson correlation coefficient, especially in the Valence and Arousal dimensions.
关键词
情绪回归 /
多任务模型 /
维度-标签信息 /
情绪分析
{{custom_keyword}} /
Key words
emotion regression /
multi-task model /
dimension-label information /
emotion analysis
{{custom_keyword}} /
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
参考文献
[1] Kim E, Klinger R.A Survey on sentiment and emotion analysis for computational literary studies[J].arXiv preprint arXiv: 1808.03137,2018.
[2] Chauhan D S, Dhanush S R, Ekbal A, et al. Sentiment and emotion help sarcasm? a multi-task learning framework for multi-modal sarcasm, sentiment and emotion analysis[C]//Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics,2020: 4351-4360.
[3] Zhu S Y, Li S S, Zhou G D. Domanial and dimensional adversarial learning for emotion regression[J].Neurocomputing,2021,420: 281-289.
[4] Khare A, Parthasarathy S, Sundaram S. Multi-modal embeddings using multi -task learning for emotion recognition[C]//Proceedings of the Interspeech,2020.
[5] Li X S, Rao Y H, Xie H R, et al. Bootstrapping social emotion classification with semantically rich hybrid neural networks[J].IEEE Transactions on Affective Computing,2017,8(4): 428-442.
[6] Wang X, Wu Y, Chen X O, et al.Enhance popular music emotion regression by importing structure information[C]//Proceedings of Asia Pacific Signal and Information Processing Association Annual Summit and Conference, 2013: 1-4.
[7] Golubev A, Loukachevitch N. Use of BERT neural network models for sentiment analysis in russian[J].Automatic Documentation and Mathematical Linguistics,2021,55(1): 17-25.
[8] Jabreel M, Moreno A. A deep learning-based approach for multi-label emotion classification in tweets[J].Applied Sciences,2019,9: 1123.
[9] Yu J F, Marujo L, Jiang J, et al. Improving multi-label emotion classification via sentiment classification with dual attention transfer network[C]//Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing,2018: 1097-1102.
[10] Barrett L F. Solving the emotion paradox: Categorization and the experience of emotion[J]. Personality and Social Psychology Review,2006, 10(1): 20-46.
[11] Buechel S, Hahn U. EMOBANK: Studying the impact of annotation perspective and representation format on dimensional emotion analysis[C]//Proceedings of the 15th Conference of the European Chapter of the Association for Computational Linguistics,2017: 578-585.
[12] Zhu S Y, Li S S, Zhou G D. Adversarial attention modeling for multi-dimensional emotion regression[C]//Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, 2019: 471-480.
[13] Rao Y H, Xie H R, Li J Y, et al.Social emotion classification of short text via topic-level maximum entropy model[J].Inf.Manag,2016,53: 978 -986.
[14] Reddy A A, Srirangam V K, Suhas D, et al.Creation of corpus and analysis in code-mixed kannada-english twitter data for emotion prediction[C]//Proceedings of the 28th International Conference on Computational Linguistic,2020.
[15] hman E, Pàmies M, Kajava K, et al. XED: A multilingual dataset for sentiment analysis and emotion detection[C]//Proceedings for the 28th International Conference on Computational Linguistics,2020: 6542-6552.
[16] Poria S, Hazarika D, Majumder N, et al.MELD: A multimodal multi-party dataset for emotion recognition in conversations[C]//Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, 2019: 527-536.
[17] He H H, Xia R. Joint binary neural network for multi-label learning with applications to emotion classification[J].arXiv preprint arXiv: 1802.00891,2018.
[18] Quan M, Tepper O, Small K, et al. Sentence emotion analysis and recognition based on emotion words using Ren-CECps[J]. International Journal of Advanced Intelligence,2010, 2(1): 105-117.
[19] Wang J, Yu L, Lai K, et al. Dimensional sentiment analysis using a regional CNN-LSTM model[C]//Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics, 2016: 225-230.
[20] Devlin J, Chang M, Lee L, et al.BERT: Pre-training of deep bidirectional transformers for language understanding.[C]//Proceedings of NAACL-HLT,2019: 4171-4186.
[21] Sun C, Yang Z H, Wang L, et al. Biomedical named entity recognition using BERT in the machine reading comprehension framework[J].Journal of Biomedical Informatics,2021: 103799.
[22] Liu Y. Fine-tune BERT for extractive summariza-tion[J].arXiv preprint arXiv: 1903.10318,2019.
[23] Gupta A, Kvernadze G, Vivek S.BERT & family eatword salad: Experiments with text understanding[J].ArXiv,2021,abs/2101.03453.
[24] Sun C, Qiu X D, Xu Y G. How to fine-tune BERT for text classification?[J]. arXiv preprint arXiv: 1905.05583,2019.
[25] Vaswani A, Shazeer N, Parmar N, et al. Attention is all you need[C]//Proceedings of 31st Confer-ence on Neural Information Processing Systems,2017.
[26] Wu Y H, Schuster M, Chen Z F,et al. Google's neural machine translation system: Bridging the gap between human and machine translation[J].arXiv preprint arXiv: 1609.08144,2016.
[27] Demszky D, Movshovitz-Attias D, Ko J, et al. GoEmotions: A dataset of fine-grained emotions[C]//Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, 2020: 4040-4054.
[28] Hochreiter S, Schmidhuber J. Long Short-Term Memory[J]. Neural Computation,1997,9(8): 1735-1780.
[29] LeCun Y, Boser B, Denker J, et al. Backpropagation applied to handwritten zip code recognition[J].Neural Computation,1989,1: 541-551.
[30] Chung J, Gulcehre C, Cho K, et al. Empirical evaluation of gated recurrent neural networks on sequence modeling[J].arXiv preprint arXiv: 1412.3555,2014.
[31] Yang Z L, Dai Z H, Yang Y M, et al. XLNet: generalized autoregressive pretraining for language understanding[C]//Proceedings of the 33rd Conference on Neural Information Processing Systems,2019.
[32] Akhtar S, Ghosal D, Ekbal A, et al. All-in-one: Emotion, sentiment and intensity prediction using a multi-task ensemble framework[J].IEEE Transactions on Affective Computing,2020: 1-7.
{{custom_fnGroup.title_cn}}
脚注
{{custom_fn.content}}
基金
人工智能应急项目(61751206);国家重点研发计划子课题(2020AAA0108604);国家自然科学基金(62106166)
{{custom_fund}}
PDF(1661 KB)
