Generating Responses with a Specific Emotion in Dialog

Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, pages 3685–3695 Florence, Italy, July 28 - August 2, 2019. c⃝2019 Association for Computational Linguistics 3685 Generating Responses with a Specific Emotion in Dialog Zhenqiao Song1,2, Xiaoqing Zheng∗1,2, Lu Liu1,2, Mu Xu3 and Xuanjing Huang1,2 1School of Computer Science, Fudan University, Shanghai, China 2Shanghai Key Laboratory of Intelligent Information Processing 3Department of Computer Science, University of California, Santa Barbara {zqsong17, zhengxq, l liu15}@fudan.edu.cn muxu@ucsb.edu, xjhuang@fudan.edu.cn Abstract It is desirable for dialog systems to have ca- pability to express specific emotions during a conversation, which has a direct, quantifiable impact on improvement of their usability and user satisfaction. After a careful investigation of real-life conversation data, we found that there are at least two ways to express emotions with language. One is to describe emotional states by explicitly using strong emotional words; another is to increase the intensity of the emotional experiences by implicitly com- bining neutral words in distinct ways. We pro- pose an emotional dialogue system (EmoDS) that can generate the meaningful responses with a coherent structure for a post, and mean- while express the desired emotion explicitly or implicitly within a unified framework. Experi- mental results showed EmoDS performed bet- ter than the baselines in BLEU, diversity and the quality of emotional expression. 1 Introduction Humans have the unique capacity to perceive com- plex, nuanced emotions, and also have the unique capability to communicate those experiences to one another with language. Although recent stud- ies (Partala and Surakka, 2004; Prendinger and Ishizuka, 2005) provide much evidence that the systems capable of expressing emotions signifi- cantly improve the user satisfaction, it is still a great challenge to make dialogue systems more “emotional” in their responses. In early representative work (Polzin and Waibel, 2000; Skowron, 2010), manually prepared rules are applied to deliberately select the desired “emo- tional” responses from a conversation corpus. Those rules were written by persons with exper- tise after careful investigation in the corpus, which makes it hard to express complex, various emo- tions, and difficult to scale well to large datasets. Post: I bought a beautiful dress yesterday! Explicit: Wearing beautiful dress makes me happy! Implicit: Wow, you must feel walking on air! Post: The rose is really beautiful! Explicit: I love rose! Implicit: I am keen on rose. Post: I lost my computer today! Explicit: It is really an annoying thing. Implicit: Oh, you must feel hot under the collar. Table 1: Examples of two (explicit and implicit) ways in emotional expressions. For each post, one emotional response for each way is listed below. The emotional words associated with strong feelings are highlighted in bold blue font. Most recently, a sequence to sequence (seq2seq) learning framework with recurrent neural net- works (RNNs) has been successfully used to build conversational agents (also known as chatbots) (Sutskever et al., 2014; Sordoni et al., 2015; Shang et al., 2015; Vinyals and Le, 2015; Serban et al., 2016a,b; Wen et al., 2016; Li et al., 2017; …

Content truncated to 500 words