Back-end semantics for multimodal dialog on XR devices

Abstract

Extended Reality (XR) devices have great potential to become the next wave in mobile interaction. They provide powerful, easy-to-use Augmented Reality (AR) and/or Mixed Reality (MR) in conjunction with multimodal interaction facilities using gaze, gesture, and speech. However, current implementations typically lack a coherent semantic representation for the virtual elements, backend-communication, and dialog capabilities. Existing devices are often restricted to mere command and control interactions. To improve these shortcomings and realize enhanced system capabilities and comprehensive interactivity, we have developed a flexible modular approach that integrates powerful back-end platforms using standard API interfaces. As a concrete example, we present our distributed implementation of a multimodal dialog system on the Microsoft Hololens®. It uses the SiAM-dp multimodal dialog platform as a back-end service and an Open Semantic Framework (OSF) back-end server to extract the semantic models for creating the dialog domain model.