Mobile spoken dialogue system using parser dependencies and ontology

In this paper, we propose a mobile spoken dialogue system with a new spoken dialogue understanding architecture (SLU). This new SLU module combines an ontology and a dependency graph to do semantic analysis. The turn analysis algorithm integrated in the SLU module uses, at each turn of the dialogue, the dependencies generated by Stanford parser and a domain ontology to analyze the sentence and to extract user’s intention and slots values (i.e. user dialogue acts, concepts and their values). The SLU module maps the sentence into a structure. The dialogue manager receives this mapping structure from the SLU module and chooses the action to be taken by the system. The mobile spoken dialogue system was developed as a remote system on a mobile phone. It utilizes Google server for recognition and Google text to speech for speech synthesis provided with Android system. Dialogue understanding, dialogue managing and text generating modules reside on a remote computer. Ten users have tested the first version of our system and a score of 3.6 on Likert scale was obtained.