A dialogue system (or a conversational agent) is a computer system that can have a dialogue with a human user.
A dialogue system consists of at least three modules: a module for the production of synthetic speech, a module for the recognition of human speech and a module for the dialogue management. The lectures will treat the speech technology components as "black boxes" and spend most of the time on theories behind dialogue management and how to combine components into usable systems. These theories will also be connected to theories of human-human dialogue and to concepts such as dialogue structure, turn-taking and feedback.
The course lectures will also discuss important standards in the field such as the W3C recommendations "Speech Recognition Grammar Specification" (SRGS), "Semantic Interpretation for Speech Recognition" (SISR), "Speech Synthesis Markup Language" (SSML) and "VoiceXML", "State Chart XML" (SCXML). Time will also be spent on introducing good and proven methodology to design dialogue systems.
The course is primarily a laboratory course and participants, assisted by their teacher, are expected to spend much of their time experimenting with speech synthesis/speech recognition and to design, implement and evaluate simple dialogue systems.
The programme is given at the Department of Philosophy, Linguistics and Theory of Science (FLoV). Link to pages about the department.