Embodied Conversational Agents
Embodied Conversational Agents (also known as ECA) are animated software agents (usually very similar to the human figure) which represent the possibility of extending the aim of common conversation systems and perform non appropriate verbal behavior, such as facial expressions, looks, gestures and processing of natural language in all its levels, contributing to a better understanding of a message. Bickmore (2003) defines such agents as entities which use, in real time, speech, gestures, observation (among other verbal and non-verbal channels) to simulate face to face human interaction.
Affective Computing techniques have been employed and investigated in order to feel people’s affection through a large variety of psychological aspects, verbal and non-verbal channels and to develop systems which demonstrate affection by making use of a variety of types, including speech, facial expressions, among others in Embodied
Conversational Agents through different modalities of communication.