ABSTRACT
Building an empathic conversation machine in open-domain is a promising research topic in natural language processing. However, most current approaches rely on designated emotions to conduct generating responses and lack the ability to decide the appropriate emotion strategy. In this paper, we propose a dialogue model of jointly predicting and generating emotions called DRCVAE, which stands for Decoupled Representations of Conditional Variational Autoencoders.
More specifically, the model separates the latent variable in conditional variational autoencoders (CVAE) into two parts: emotion and content. Then the latent emotional strategy (implicit) is further forced to predict the target emotion probability distribution (explicit). By using implicit and explicit emotional strategy, a newly designed paired decoder incorporates rich control information to decode the response. Experiment results demonstrate that DRCVAE provides an effective way to infer target emotions and generate high-quality responses simultaneously.
- Mayer, J. D., & Geher, G. (1996). Emotional intelligence and the identification of emotion. Intelligence, 22(2), 89--113.Google ScholarCross Ref
- Bowman, S. R., Vilnis, L., Vinyals, O., Dai, A. M., Jozefowicz, R., & Bengio, S. (2015). Generating sentences from a continuous space. arXiv preprint arXiv:1511.06349.Google Scholar
- Chung, J., Gulcehre, C., Cho, K., & Bengio, Y. (2014). Empirical evaluation of gated recurrent neural networks on sequence modeling. arXiv preprint arXiv:1412.3555.Google ScholarDigital Library
- Hochreiter, S., & Schmidhuber, J. (1997). Long short-term memory. Neural computation, 9(8), 1735--1780.Google Scholar
- Hu, X., Tang, J., Gao, H., & Liu, H. (2013, May). Unsupervised sentiment analysis with emotional signals. In Proceedings of the 22nd international conference on World Wide Web (pp. 607--618). ACM.Google ScholarDigital Library
- Jiang, L., Yu, M., Zhou, M., Liu, X., & Zhao, T. (2011, June). Target-dependent twitter sentiment classification. In Proceedings of the 49th Annual Meeting of the Association for Computational Linguistics: Human Language Technologies-Volume 1 (pp. 151--160). Association for Computational Linguistics.Google ScholarDigital Library
- Kingma, D. P., & Ba, J. (2014). Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980.Google Scholar
- Kingma, D. P., & Welling, M. (2013). Autoencoding variational bayes. arXiv preprint arXiv:1312.6114.Google Scholar
- Krizhevsky, A., Sutskever, I., & Hinton, G. E. (2012). Imagenet classification with deep convolutional neural networks. In Advances in neural information processing systems (pp. 1097--1105).Google ScholarDigital Library
- Larsen, A. B. L., Sønderby, S. K., Larochelle, H., & Winther, O. (2015). Autoencoding beyond pixels using a learned similarity metric. arXiv preprint arXiv:1512.09300.Google Scholar
- Moore, B. (1981). Principal component analysis in linear systems: Controllability, observability, and model reduction. IEEE transactions on automatic control, 26(1), 17--32.Google Scholar
- Pang, B., Lee, L., & Vaithyanathan, S. (2002, July). Thumbs up? sentiment classification using machine learning techniques. In Proceedings of the ACL-02 conference on Empirical methods in natural language processing-Volume 10(pp. 79--86). Association for Computational Linguistics.Google Scholar
- Rezende, D. J., Mohamed, S., & Wierstra, D. (2014). Stochastic backpropagation and approximate inference in deep generative models. arXiv preprint arXiv:1401.4082.Google Scholar
- Schuster, M., & Paliwal, K. K. (1997). Bidirectional recurrent neural networks. IEEE Transactions on Signal Processing, 45(11), 2673--2681.Google ScholarDigital Library
- Serban, I. V., Sordoni, A., Bengio, Y., Courville, A. C., & Pineau, J. (2016, February). Building End-To-End Dialogue Systems Using Generative Hierarchical Neural Network Models. In AAAI (Vol. 16, pp. 3776--3784).Google Scholar
- Shang, L., Lu, Z., & Li, H. (2015). Neural responding machine for short-text conversation. arXiv preprint arXiv:1503.02364.Google Scholar
- Sohn, K., Lee, H., & Yan, X. (2015). Learning structured output representation using deep conditional generative models. In Advances in Neural Information Processing Systems (pp. 3483--3491).Google Scholar
- Sun, X., Zhang, C., & Li, L. (2019). Dynamic emotion modelling and anomaly detection in conversation based on emotional transition tensor. Information Fusion, 46, 11--22.Google ScholarDigital Library
- Tallec, C., & Ollivier, Y. (2018). Can recurrent neural networks warp time? arXiv preprint arXiv:1804.11188.Google Scholar
- Wawre, S. V., & Deshmukh, S. N. (2016). Sentiment classification using machine learning techniques. International Journal of Science and Research (IJSR), 5(4), 819--821.Google ScholarCross Ref
- Wen, T. H., Miao, Y., Blunsom, P., & Young, S. (2017, July). Latent Intention Dialogue Models. In International Conference on Machine Learning (pp. 3732--3741).Google Scholar
- van der Westhuizen, J., & Lasenby, J. (2018). The unreasonable effectiveness of the forget gate. arXiv preprint arXiv:1804.04849.Google Scholar
- Xu, Linhong, et al. "Constructing the affective lexicon ontology." Journal of the China Society for Scientific and Technical Information 27.2 (2008): 180--185.Google Scholar
- Zhao, Tiancheng, Ran Zhao, and Maxine Eskenazi. "Learning Discourse-level Diversity for Neural Dialog Models using Conditional Variational Autoencoders." Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). Vol. 1. 2017.Google Scholar
- Zhou, Hao, et al. "Emotional chatting machine: Emotional conversation generation with internal and external memory." arXiv preprint arXiv:1704.01074 (2017).Google Scholar
- Zhou, X., & Wang, W. Y. MOJITALK: Generating Emotional Responses at Scale (Supplementary Materials).Google Scholar
- Asghar, N., Poupart, P., Hoey, J., Jiang, X., & Mou, L. (2018, March). Affective Neural Response Generation. In European Conference on Information Retrieval (pp. 154--166). Springer, Cham.Google Scholar
Index Terms
- Towards Automated Emotional Conversation Generation with Implicit and Explicit Affective Strategy
Recommendations
Neural Conversation Generation with Auxiliary Emotional Supervised Models
An important aspect of developing dialogue agents involves endowing a conversation system with emotion perception and interaction. Most existing emotion dialogue models lack the adaptability and extensibility of different scenes because of their ...
Neural Emotional Response Generation via Adversarial Transfer Learning
ICIAI '19: Proceedings of the 2019 3rd International Conference on Innovation in Artificial IntelligenceEmotional response generation is a key step to build an empathetic chatbot. However, previous emotional chatting models mainly focus on single-turn conversation, and multi-turn context emotional response generation has not been explored. In this paper, ...
Mediating individual affective experience through the emotional photo frame
Affect Aware Ubiquitous ComputingA photograph is considered a medium with emotional legibility and a means of expressing and exchanging emotional experience. This research presents the interactive emotional photo frame system focusing on mediating individual affective experience among ...
Comments