Speech interface for controlling an hi-fi audio system based on a Bayesian belief networks approach for dialog modeling

Abstract

This paper presents the development of a speech interface for controlling a high fidelity system from natural language sentences. A Bayesian Belief Network approach is proposed for dialog modeling. This solution is applied to infer the user's goals corresponding to the processed utterances. Subsequently, from the inferred goals, missing or spurious concepts are automatically detected. This is used to drive the dialog prompting for missing concepts and clarifying for spurious concepts allowing more flexible and natural dialogs. A dialog strategy which makes use of the dialog history and the system's state is also presented.

Contributions to the contextualization of human-machine spoken interaction systems

“…Figure 2.1 presents, as an example of a SDS, our proprietary system, that we will use as a baseline for the work proposed in this Thesis. A more comprehensive study of this system can be consulted in Fernández et al (2005); Fernández-Martínez (2009).…”

“…That is, the information provided by the user in his or her utterance, along with the information gathered in previous turns. An additional inference process is then conducted to determine whether the system has all the information it needs to carry out the inferred goals (Fernández et al, 2005). We will see in Chapter 5 that this scheme endows the system with more flexibility, since it can move from one goal to another without the need of finishing the first one, therefore giving place to more natural dialogues.…”

“…Among the guidance systems we could mention several initiatives to interact with a robot using speech, such as the project HYGEIOROBOT, for a health environment (Spiliotopoulos et al, 2001), or the project URBANO and its following projects, ROBINT and ROBONAUTA, for offering tour guides in a Science museum (Lucas et al, 2008;Rodríguez-Losada et al, 2008). Finally, as control systems we could mention the applications to control household devices, such as a television (Oh et al, 2007a) or an audio system (Fernández et al, 2005), or in-car dialogue systems to control aspects such as navigation, climate or entertainment (Hassel and Hagen, 2006).…”

“…the dialogue goals), and the information extracted by the understanding module (i.e. dialogue concepts), that are needed to achieve those actions, whether available or not (Fernández et al (2005)).…”

Contributions to speech analytics based on speech recognition and topic identification

“…The main purpose of a dialogue system can be very diverse, however in accordance with McTear [2002] it is worth highlighting the dialogue systems applied to the retrieval of information, services and transactions. Examples of these systems are services for booking train tickets by phone , information and recommendation systems for movies [Chu-Carroll, 2000], and control systems for household appliances [Fernández et al, 2005.…”