David Duff and Susann Luperfoy
We have constructed the discourse component for a multimodal user-interface dialogue system. The larger system acts as a conversational agent engaged in dialogue with the human user. Output originating with the backend application system can take the eventual form of spoken English utterances, printed English text, or changes in the graphical display screen, e.g., appearance, disappearance, movement, or color change of a screen icon, and left or right panning or zooming of the overall display. User input utterances can be typed or spoken English queries and commands, mouse gestures (simple clicks), and typed English command-language utterances.