Computational Model of Listener Behavior for Embodied Conversational Agents (Google eBook) 0 Reviewshttp://books.google.com/books/about/Computational_Model_of_Listener_Behavior.html?id=SvFbyHyBPYMC Universal-Publishers, 2009 - 166 pages During an interaction, listeners do not assimilate passively all the speaker words, but they actively participate to the interaction providing information about how they feel and what they think of the speaker's speech. According to the listener's behavior, the speaker can estimate how his interlocutor is reacting and can decide how to carry on the interaction. In this thesis, to refer to signals provided by listeners, we adopt the term backchannel and we define backchannels as acoustic and non-verbal signals provided during the speaker's turn to exchange information about the communicative functions: contact, perception, understanding and attitude. Two fundamental characteristics of backchannels are : (i) they can be emitted at different level of intentionality; (ii) they can be "reactive" (deriving from a first process of perception of the speaker's speech) or "response" (deriving from a more aware evaluation). A particular form of backchannel is the mimicry of the speaker's behavior: that is the behavior displayed by an individual who does what another person does. This type of behavior has been proven to play quite an important role during conversations. Due to the importance of the listener's behavior, in this thesis we propose to implement a model that generates this type of behavior for an ECA while interacting with a user. We aim to improve the human-machine interaction. Our contributions are: (i) a repository of backchannel signals; (ii) a real-time system able to generate the types of backchannel "reactive", "response" and "mimicry" at a low level of intentionality; (iii) an evaluation of the system through perceptive tests and applications. |
