tua1 at the ntcir 14 stc 3 task
play

TUA1 at the NTCIR-14 STC-3 Task Chinese Emotional Conversation - PowerPoint PPT Presentation

TUA1 at the NTCIR-14 STC-3 Task Chinese Emotional Conversation Generation Subtask Tokushima University Department of Information Science & Intelligent Systems Yangyang Zhou, Zheng Liu, Xin Kang, Yunong Wu, and Fuji Ren Faculty of


  1. TUA1 at the NTCIR-14 STC-3 Task Chinese Emotional Conversation Generation Subtask Tokushima University Department of Information Science & Intelligent Systems Yangyang Zhou, Zheng Liu, Xin Kang, Yunong Wu, and Fuji Ren Faculty of Engineering the Department of Information Science & Intelligent Systems The University of Tokushima, Ren Lab

  2. Ren Contents Lab  Background  Purpose  Related work  Proposed method  Data processing  Experiment  Evaluation  Conclusion and future work  References 2 Faculty of Engineering the Department of Information Science & Intelligent Systems The University of Tokushima, Ren Lab Faculty of Engineering the Department of Information Science & Intelligent Systems The University of Tokushima, Ren Lab

  3. Ren Background Lab More appropriate response Can I love you? What love do you talk about? Kneeling. You deserve to look good all your life. You're bragging about yourself again. Turing test You are 1.4 meters today. 3 Faculty of Engineering the Department of Information Science & Intelligent Systems The University of Tokushima, Ren Lab Faculty of Engineering the Department of Information Science & Intelligent Systems The University of Tokushima, Ren Lab

  4. Ren Purpose Lab Post Emotion Category Response (Given) (Given) (to be generated) 爱狗还会做饭的男人,最帅了! 会做饭的男人是很帅的啊。 喜欢 The man who cooks and The man who cooks is Like loves dogs is very handsome! handsome. 本来想学一把沧桑,结果令我 更忧伤。 悲伤 这是一个悲伤的故事。 I wanted to learn the Sadness It a sad story. vicissitudes of life, but I became sadder. 6 categories : {Anger, Disgust, Happiness, Like, Sadness, Other} 4 Faculty of Engineering the Department of Information Science & Intelligent Systems The University of Tokushima, Ren Lab Faculty of Engineering the Department of Information Science & Intelligent Systems The University of Tokushima, Ren Lab

  5. Ren Related work Lab Sequence to sequence problem Chinese emotion analysis and recognition application 5 Faculty of Engineering the Department of Information Science & Intelligent Systems The University of Tokushima, Ren Lab Faculty of Engineering the Department of Information Science & Intelligent Systems The University of Tokushima, Ren Lab

  6. Ren Proposed method Lab P&E2R 爱狗还会做饭的男人,最帅了! 会做饭的男人是很帅的啊。 喜欢 The man who cooks and loves The man who cooks is Like dogs is very handsome! handsome. 爱狗还会做饭的男人,最帅了! + 喜欢 + <start> 会 爱狗还会做饭的男人,最帅了! + 喜欢 + 会 做 …… …… …… 爱狗还会做饭的男人,最帅了! + 喜欢 + 。 6 <end> Faculty of Engineering the Department of Information Science & Intelligent Systems The University of Tokushima, Ren Lab Faculty of Engineering the Department of Information Science & Intelligent Systems The University of Tokushima, Ren Lab

  7. Ren Proposed method Lab P2R&E2R 爱狗还会做饭的男人,最帅了! 会做饭的男人是很帅的啊。 喜欢 The man who cooks and loves The man who cooks is Like dogs is very handsome! handsome. 爱狗还会做饭的男人,最帅了! + <start> 会 喜欢 + <start> 爱狗还会做饭的男人,最帅了! + 会 做 喜欢 + 会 …… …… …… 爱狗还会做饭的男人,最帅了! + 。 7 <end> 喜欢 + 。 Faculty of Engineering the Department of Information Science & Intelligent Systems The University of Tokushima, Ren Lab Faculty of Engineering the Department of Information Science & Intelligent Systems The University of Tokushima, Ren Lab

  8. Ren Data processing Lab Over 1.7 million Weibo post-response pairs • Removing pairs without Chinese characters. • e.g. - How are you? 1 – Fine. • Removing extra duplicate characters (3 times at most). • e.g. – 哈哈哈哈 2 哈!!!! • Removing low- frequent (frequency Distribution of posts & responses length < 50) characters. 0.32% of responses are longer than 30+2 characters 3 • e.g. – 这是饕餮。 Extra dataset 40 thousand sentences and corresponding emotion labels 8 Faculty of Engineering the Department of Information Science & Intelligent Systems The University of Tokushima, Ren Lab Faculty of Engineering the Department of Information Science & Intelligent Systems The University of Tokushima, Ren Lab

  9. Ren Evaluation Lab 200 posts x 5 emotions Given post: 爱狗还会做饭的男人,最帅了! The man who cooks and loves dogs is very handsome! Chinese/Translated English Emotion Coherence Emotion Response Consistency Label Class and Fluency 会做饭的男人是很帅的啊。 喜欢 Response 1 The man who cooks is Yes Yes 2 Like handsome. 哈哈,我也觉得。 喜欢 Response 2 Yes No 1 Haha, I feel the same way. Like 这是哪部电影里的? 厌恶 Response 3 No Yes 0 Which movie is this from? Disgust 哈哈 , 你也是。 喜欢 Response 4 No No 0 Haha, the same to you. Like 9 Faculty of Engineering the Department of Information Science & Intelligent Systems The University of Tokushima, Ren Lab Faculty of Engineering the Department of Information Science & Intelligent Systems The University of Tokushima, Ren Lab

  10. Ren Experiment Lab Evaluation results of our run submissions 10 Faculty of Engineering the Department of Information Science & Intelligent Systems The University of Tokushima, Ren Lab Faculty of Engineering the Department of Information Science & Intelligent Systems The University of Tokushima, Ren Lab

  11. Ren Conclusion and future work Lab Data • Generate Result processing emotional responses • P&E2R • Average • 2 datasets by given • P2R&E2R scores > 0.8 posts • 3 removing Purpose Method Future work • Diversity of responses 11 Faculty of Engineering the Department of Information Science & Intelligent Systems The University of Tokushima, Ren Lab Faculty of Engineering the Department of Information Science & Intelligent Systems The University of Tokushima, Ren Lab

  12. Ren References Lab 1. Sutskever, I., Vinyals, O., Le, Q.V.: Sequence to sequence learning with neural networks. In: Advances in neural information processing systems. pp. 3104 – 3112 (2014) 2. Vijayakumar, A.K., Cogswell, M., Selvaraju, R.R., Sun, Q., Lee, S., Crandall, D., Batra, D.: Diverse beam search: Decoding diverse solutions from neural sequence models. arXiv preprint arXiv:1610.02424 (2016) 3. Ghosh, S., Chollet, M., Laksana, E., Morency , L.P., Scherer, S.: Affect -lm: A neural language model for customizable affective text generation. arXiv preprint arXiv:1704.06851 (2017) 4. Zhou, H., Huang, M., Zhang, T., Zhu, X., Liu, B.: Emotional chatting machine: Emotional conversation generation with internal and external memory. In: ThirtySecond AAAI Conference on Artifcial Intelligence (2018) 5. Gao, F., Sun, X., Wang, K., Ren, F.: Chinese micro-blog sentiment analysis based on semantic features and pad model. In: 2016 IEEE/ACIS 15th International Conference on Computer and Information Science (ICIS). pp. 1 – 5. IEEE (2016) 6. Quan, C., Ren, F.: Visualizing emotions from chinese blogs by textual emotion analysis and recognition techniques. International Journal of Information Technology & Decision Making 15(01), 215 – 234 (2016) 12 Faculty of Engineering the Department of Information Science & Intelligent Systems The University of Tokushima, Ren Lab Faculty of Engineering the Department of Information Science & Intelligent Systems The University of Tokushima, Ren Lab

Recommend


More recommend