Offensive Language Identification in Greek

Proceedings of the 12th Conference on Language Resources and Evaluation (LREC 2020), pages 491–497 Marseille, 11–16 May 2020 c⃝ European Language Resources Association (ELRA), licensed under CC-BY-NC 491 Emotional Speech Corpus for Persuasive Dialogue System Sara Asai1, Koichiro Yoshino1,2,3, Seitaro Shinagawa1,3, Sakriani Sakti1,3, Satoshi Nakamura1,3 1Nara Institute of Science and Technology, Takayama 8916-5, Ikoma, Nara, 6300192, Japan 2PRESTO, Japan Science and Technology Agency, Saitama, Japan 3RIKEN Center for Advanced Intelligence Project (AIP), Tokyo, Japan {koichiro, shinagawa.seitaro.si8, ssakti, s-nakamura} at is.naist.jp Abstract Expressing emotion is known as an efficient way to persuade one’s dialogue partner to accept one’s claim or proposal. Emotional expression in speech can express the speaker’s emotion more directly than using only emotion expression in text, which will lead to a more persuasive dialogue. In this paper, we built a speech dialogue corpus in a persuasive scenario that uses emotional expressions to build a persuasive dialogue system with emotional expressions. We extended an existing text dialogue corpus by adding variations of emotional responses to cover different combinations of broad dialogue context and a variety of emotional states by crowd-sourcing. Then, we recorded emotional speech consisting of collected emotional expressions spoken by a voice actor. The experimental results indicate that the collected emotional expressions with their speeches have higher emotional expressiveness for expressing the system’s emotion to users. Keywords: dialogue corpus, emotional expression, speech corpus, persuasive dialogue 1. Introduction Persuasion or negotiation is an important dia- logue style, which has been widely researched recently (Mazzotta et al., 2007; Georgila, 2013; Hiraoka et al., 2016; Wang et al., 2019). Emotional expressions have an important role in various dia- logue situations and contexts (Keltner and Haidt, 1999; Morris and Keltner, 2000; Adler et al., 2016). It is well known that emotional expressions are useful in persuasion and negotiation (Fogg, 1999): building a cooperative relationship with positive expressions (Forgas, 1998) or pressing the dialogue partner to accept a proposal with negative expressions (Sinaceur and Tiedens, 2006). We built dialogue corpora in a persuasive scenario annotated with emotion labels to build persuasive dialogue systems that can use emotional expressions to improve its success rate (Yoshino et al., 2018). When the system uses emotional expression in a dialogue, it is important to correctly express the emotion that the sys- tem intended. Expressing actual emotion to the users with only textual information is sometimes difficult because tex- tual information has limited expressiveness. In contrast, emotional speech or gesture has the potential to improve the expressiveness for expressing the intended emotional state to the user. In this paper, we collected emotional expressions for the persuasive scenario and recorded their audio by expressing the emotional state to be indicated to the dialogue partner. Existing dialogue corpora based on natural scenario collec- tion (Yoshino et al., 2018) do not have comprehensive emo- tional expressions for any dialogue contexts. However, for building a dialogue system that can use any emotional states in any dialogue contexts, we collected dialogue responses based on any emotional states given a dialogue context via …

Content truncated to 500 words