Spoken-dialogue Interaction for Serious Games and Virtual Training Simulators
(Room S320D)
26 Nov 18
12:45 PM
-
2:15 PM
Recent advances in deep learning have opened people’s minds to new levels of interaction within many Virtual Training Environments (VTEs)/Serious Games. Unfortunately, allowing users to fully interact through speech is a challenge for computers, particularly when the task is unconstrained or performed under adverse conditions. Speech is therefore often neglected as a modality, in spite of its naturalness.
Furthermore, user-based evaluations of speech interfaces are intrinsically difficult. The recent AI revolution brings hope that, despite these shortcomings, there will at least be some areas of R&D that will lead to real improvements in both the design and implementation of spoken interactions in training systems. This tutorial will explain: how deep learning has transformed dialogue/chat systems; the components of a typical dialogue system; the challenges in enabling spoken dialogue as a modality for hands-free interaction; the major usability issues for AI/dialogue interaction within VTEs/serious games; opportunities for researchers and developers to enhance system interactivity through AI/dialogue; and how to enable dialogue interaction within immersive, mixed-reality environments.