104 The use of language in affective computing for human-computer interaction

The field of Human-Computer Interaction has traditionally focused on the study, design, and implementation of techniques that facilitate an effective communication between a computational system and the user. In the last decades, the developments in this area have evolved from the design of useful and usable interfaces to the construction of complex interactive systems able to motivate and engage the user in the execution of a specific task. A relatively new field of research that has contributed to the advancement of human-computer interaction in this direction is known as affective computing. The term affective computing was coined in 1997 to refer to efforts dedicated to the construction of artificial systems able to recognise emotional states from the user and simulate emotions in advanced interactive systems such as embodied conversational agents and robots. A key component in the development of these systems is the use of natural language, as a source to identify human emotions, and as a medium to interact with the user. In this chapter we present a brief overview of the affective computing research area, emphasising how language is used for emotion recognition and emotional speech production. In the second part of the chapter, we describe our work in the recognition of emotions using paralinguistic information and the modelling of emotions for embodied conversational agents.