An Analysis of the Effect of Emotional Speech Synthesis on Non-Task-Oriented Dialogue System

Yuya Chiba,Taketo Kase,Mai Yamanaka,Akinori Ito,Takashi Nose

doi:10.18653/v1/w18-5044

Abstract

This paper explores the effect of emotional speech synthesis on a spoken dialogue system when the dialogue is non-task-oriented. Although the use of emotional speech responses have been shown to be effective in a limited domain, e.g., scenario-based and counseling dialogue, the effect is still not clear in the non-task-oriented dialogue such as voice chatting. For this purpose, we constructed a simple dialogue system with example- and rule-based dialogue management. In the system, two types of emotion labeling with emotion estimation are adopted, i.e., system-driven and user-cooperative emotion labeling. We conducted a dialogue experiment where subjects evaluate the subjective quality of the system and the dialogue from the multiple aspects such as richness of the dialogue and impression of the agent. We then analyze and discuss the results and show the advantage of using appropriate emotions for the expressive speech responses in the non-task-oriented system.

Highlights

Spoken dialogue systems have been becoming popular in various applications, such as a speech assistant system in smartphones and smart speakers, an information guide system in public places, and humanoid robots
There have been a variety of studies for developing spoken dialogue systems, and the systems are roughly grouped into two categories, taskoriented and non-task-oriented systems, from the aspect of having a goal or not in the dialogue
The task-oriented dialogue systems (Zue et al, 2000; Kawanami et al, 2007) are important as practical applications, e.g., ticket vending and information guidance, the role of the non-task-oriented systems is increasing for more advanced human-computer interaction (HCI) including voice chat

Summary

Introduction

Spoken dialogue systems have been becoming popular in various applications, such as a speech assistant system in smartphones and smart speakers, an information guide system in public places, and humanoid robots. The task-oriented dialogue systems (Zue et al, 2000; Kawanami et al, 2007) are important as practical applications, e.g., ticket vending and information guidance, the role of the non-task-oriented systems is increasing for more advanced human-computer interaction (HCI) including voice chat. Yu et al (2016) proposed a set of conversational strategies to handle possible system breakdowns. Nakano et al (2006) tried to incorporate both task-oriented and non-taskoriented dialogue functions into a humanoid robot using a multi-expert model. These studies enhance the performance of the dialogue systems, an important role is still missing from the viewpoint of the system expressivity. The system cannot perceive and express para-linguistic information such as emotions, which is completely different from our daily communication

Methods

Results

Conclusion